Interactions Between Domestic, Invasive and Threatened Carnivores and Their Implications in Conservation and Pathogen Transmission

Home , Marine otter

Interactions between domestic, invasive and threatened carnivores and their implications in conservation and pathogen transmission

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY

Maximiliano A. Sepúlveda

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Randall Singer, Katey Pelican

July 2013

Acknowledgments

To write this section is a difficult and sensitive task where I have to take a look backwards and remember all those people that supported me and helped me get to this final stage: a hard and beautiful road. First at all, I want to thank my co-advisers Dr.

Katey Pelican and Dr. Randall Singer; their experience in international research gave them confidence to support my work in Chile. I thank them for their support and excellent advice in all aspects of my training: writing grants, facilitating my field work in

Chile, connecting me with excellent collaborators, and fruitful discussions during my manuscript writing. Your respective work and research inspired me to believe in the importance and impact of research in the development of a better society.

I also want to thank to Dr. Michael Stoskopf, at North Carolina State University, who was my first formal adviser and who motivated me to not only think about adequate questions but also to think in new and innovative ways to address those questions.

During my time in the USA I also had the opportunity to work with professors that impressed me, not only by their high academic achievements but also by the way they interact with their colleagues and students, approach I hope to emulate: Drs. Paul Cross

(USGS, Bozeman), George Hess (North Carolina State University), Peter Hudson

(Pennsylvania State University) and Andrew Dobson (Princeton University).

My significant thanks to Dr. Claudia Muñoz-Zanzi and Dr. Dave Garshelis, my committee members, who gave me excellent recommendations and ideas from their respective fields of experience and provided enriching and helpful advice from the first to the last committee meetings as well as important edits to this document.

In Chile, many people were critical to the success of this work, particularly considering that this dissertation relies on data collected during two years of intensive field work. I want to especially thank the great support of the Valdivian Coastal Reserve (The Nature

Conservancy) who provided their excellent facilities for our work (cabin, hot water, internet, laboratory, office space, almost like a place for vacation time!). Particular thanks go to the park rangers, Danilo Gonzalez, Erwin Ovando, Gerardo Ponce, Omar

Ponce, Patricia Poveda; and also to Alfredo Almonacid, Liliana Pezoa and Francisco

Solis from The Nature Conservancy Valdivia and Santiago.

I have been collaborating with The Comité Pro Defensa de la Flora y Fauna since 2004.

This Chilean organization gave me the first opportunity to work in the conservation and research of endangered wildlife as part of the Southern River Otter Conservation Project and also supported this project with logistics, a motored canoe, a 4 wheel drive vehicle and administrative help. Special thanks to Luz Alegría, Ana Araya, and my friend Ignacio

Rodríguez.

I am a very lucky person because of the all the good people who helped me during the many different stages of my work often under adverse and difficult conditions. Thank you so much to my colleagues and good friends: Alejandro Aleuy, Antonieta Eguren,

Angelo Espinoza, Rocio Jara, Mauricio Paredes, Tatiana Proboste, Marcela Rojas, Carlos

Sanchez, Paulina Stowhas, as well as the more than 20 excellent volunteers who helped us in the field.

Splitting the last 5 ½ years between Chile and the US has not been easy in terms of a place to call “home”. Some good friends have opened their homes to me, providing me not only with a place to stay but also with good friends and good times. Thanks to Dr.

Gerardo Acosta, Jessica Vera and Maira in Chile and Claudia Fernandez in the USA.

Thanks for the unforgettable moments.

I also want to thank to Dr. Eduardo Silva. We both share same Fulbright Fellowship cohort in 2007, with similar interests in the study of the Temperate Forest wildlife. We did our respective doctoral research in the Valdivian Coastal Reserve providing not only fascinating conversations and project collaborations, but also new discoveries such as the new population of the Darwin’s fox (with Dr. Ariel Farias). I am very fortunate to have such a good friend and brilliant scientist to share the passion of conservation and research with.

iii

Finally I want to thank to all the people from the rural communities of Curiñanco, Huape,

Chaihuín, Huiro and particularly Cadillal Alto. Cadillal Alto is a magic place with magic people where the people and the forest are one. Thanks to Don Pío and Doña Nina for your hospitality, you live in a unique paradise, isolated from our noise and beautifully protected between the mountains, a true spirit of life, love and devotion.

Dedication

To my parents Gloria and Enrique and my sister Cristina, a family of dreamers, here is my dream made real by your continuous support during all my life.

To Paulina Stowhas, my girlfriend, partner and friend. I am so happy I met you in the place and moment that I enjoy the most, the forest and the research. Now you are an important part of my life. Thank you for being such a natural and beautiful person. Every day I learn something new just by looking, watching and listening to you, beautiful. Love you a lot.

Finally, this and all my work is dedicated to my son, Alvaro, you are my connection to the earth, to the ground, and to the magic of life. I am so proud of you, and now I hope you are proud of me by achieving this, it was a huge effort, our effort. I love you so much.

To you all, you make me a fortunate and happy person every day, thanks.

Abstract

When two individuals occur at the same place and time, a contact interaction occurs. In conservation, interactions between domestic animals and/or invasive species and wildlife pose a significant threat to endangered populations due to direct conflict, conflict over resources and, possibly most concerning, risk of infectious disease. However, in most cases, these interactions are not well characterized, particularly related to the drivers of why species interact, making managing these risks challenging. This is particularly true for disease risk. For infectious pathogens, such contact provides the most critical event for their persistence: disease transmission. In carnivores, this question of transmission and persistence is even more acute. In general, carnivores have low population densities and direct interactions with other carnivores tend to be rare or avoided. How, then, do diseases persist in these populations? This dissertation investigated the factors that drive interactions between species using a multi-species carnivore model composed of domestic dogs, invasive American mink and endangered Southern river otters in the

Temperate Forest of Chile. Of particular interest as a model disease in this system was

Canine Distemper Virus (CDV), a multi-host pathogen of carnivores and a disease of high conservation concern for endangered carnivores around the world. Previous work in wildlife indicates that multi-host systems are critical to the persistence of CDV in wild carnivores, making understanding interactions in a multi-carnivore system particularly important. Using this multi carnivore system, three studies were undertaken to better characterize drivers of carnivore interactions across species: Study 1 investigated how

vi human management and use of domestic dogs influences dog-wildlife interactions using rural household surveys. Study 2 characterized domestic dog use of the landscape and risk factors for movement into protected areas and wildlife habitat. Finally, study 3 determined disease transmission risk among the three carnivore populations by characterizing interactions through household interviews and camera trapping using CDV as a model disease system. Results from Study 1 determined that dog-wild carnivore interactions are highly associated with dog’s role as a guardian of livestock and are often encouraged by owners. This is in contrast to interactions with prey species that were associated with inadequate diet and were considered undesirable. Humans also shaped rural dog’s demography dramatically, controlling their population through culling and highly skewed male: female ratios. In fact, so few animals were produced through reproduction, that most animals were imported from urban areas, providing an interesting opportunity to externally influence the population, control disease, and manage risk.

Results from Study 2, indicate that land use by dogs is mostly concentrated within 200 meters of households and exhibits a notable diurnal pattern with dogs moving mostly during the day. Characterization of movement away from the house of origin (forays) found that dogs primarily use pastures to move with fewer visitations in continuous forest habitats including those in protected areas. Rivers appeared to present a barrier for dog movement; however, there were instances of them being crossed allowing access of dogs to protected lands. When moving in forest, dogs mostly used trails and roads. Finally,

Study 3 determined that American mink interactions could potentially play a key role in

CDV transmission and possibly persistence in this carnivore assemblage. Minks

vii interacted frequently with otters at river otter latrines at intervals theoretically adequate for viable indirect transmission of CDV. As expected, dogs were rare in otter habitats; but dogs were seen interacting with American mink (killings or harassment) near households. Thus, mink in this system were acting as a potential ‘bridge host’ between dog and otter populations. In addition, a high percentage of dogs and mink had serological evidence of exposure to CDV making the potential disease transfer risk to otters real. Altogether, this dissertation provides important information about the drivers of carnivore interactions related to disease risk in a multi-carnivore system. These interactions are not random, but rather they are driven by behavior both on the part of humans and their use and management of dogs, as well as by species-specific movement and territorial behaviors. Understanding rural dog’s use and management by farmers and how dogs move on the landscape will also provide useful information to improve the integrated management of dog-wildlife interactions around protected lands. This will be important for the management not only of direct and indirect conflict, but also of disease risk in an apparent CDV reservoir in this natural ecosystem. Particularly interesting, the recent invasion of the American mink further complicates interaction risk as their movements on the landscape alter the carnivore community composition and, therefore, interactions among populations. This new population of mink may even provide sufficient population density in this ecosystem to allow persistence of CDV and other diseases that would otherwise outstrip its host population and collapse. This question of how introduced species alter population interactions in the context of disease transmission risk is poorly understood and bears future research. Altogether, the work

viii presented in this dissertation shows the importance of incorporating human drivers in the management of conservation areas. In this system, understanding human management and use of dogs, dog movements on the landscape, and how human-introduced invasive species are changing multi-host pathogen dynamics are all critical factors in protecting threatened carnivores in the Valdivian Temperate Forest.

Table of contents

List of Tables …………………………………………………………………………...xii

List of Figures ………………………………..………………………………………..xiii

CHAPTER 1: General introduction ...... 1

Canine distemper virus ...... 2

Distemper transmission and pathogenesis ...... 5

CDV as a risk to endangered populations ...... 7

Carnivore interactions ...... 13

Framing the research problem ...... 16

CHAPTER 2: Domestic dogs in rural communities around protected areas: conservation problem or conflict solution? ...... 18

Introduction ...... 18

Methods...... 20

Results ...... 25

Discussion ...... 28

Supplementary material ...... 42

CHAPTER 3: Fine-scale movements of rural domestic dogs in sensitive conservation areas in the South American Temperate Forest ...... 49

Introduction ...... 49

Methods...... 52

Results ...... 58

Discussion ...... 60

Suplementary material ...... 75

CHAPTER 4: Invasive American mink: linking pathogen risk between domestic and endangered carnivores ...... 76

Introduction ...... 76

Methods...... 83

Results ...... 89

Discussion ...... 91

Suplementary material ...... 109

CHAPTER 5: General conclusions ...... 113

Conservation management implications ...... 119

Future research recommendations ...... 122

REFERENCES ...... 125

List of Tables

Chapter 2

Table 1. Demography and management of domestic dogs in rural areas around three

protected areas in Southern Chile...... 35

Table 2. Number of households with dogs (n = 123) indicating dog-wildlife

interactions observed during the previous year...... 36

Table 3. Summary of model selection to estimate the probability of owned dog

interactions with carnivores and prey species...... 37

Chapter 3

Table 1. Summary of foray movements of 14 radio-collared rural dogs around the

Valdivian Coastal Reserve and Alerce Costero National Park...... 67

Table 2. Habitat use of dog forays at second order and third order of selection...... 68

Chapter 4

Table 1. Multivariable logistic regression results of risk factors associated with

CDV antibodies owned dogs (n=60) in Chaihuín River, Chile...... 100

Table 2. Multivariable negative binomial regression to estimate the Incidence Rate

Ratio of American mink visits to a camera trap site...... 101

Table 3 . Time intervals of different directional interaction between otter, minks and

dogs detected at camera traps...... 102

xii

List of Figures

Chapter 2

Figure 1. Map of the study area in the Coastal Range of southern Chile...... 38

Figure 2. Farm animal ownership in rural households around protected areas...... 39

Figure 3. Measures of farm animal protection against carnivore predators...... 40

Figure 4. Model-averaged parameter estimates of rural dog interactions with

carnivore and prey species...... 41

Chapter 3

Figure 1. Study area with the households that included radiotelemetred dogs...... 69

Figure 2. Track of a rural domestic dog and foray dog definition...... 70

Figure 3. Percentage of total locations by distances from dog owner’s households

per dog...... 71

Figure 4. Proportion of habitat composition for total study area, foray ranges and

radiolocations by dogs...... 72

Figure 5. Total GPS locations of 14 rural domestic dogs monitored in southern

Chile...... 73

Figure 6. Photos of rural owned dogs ranging in the study area...... 74

Chapter 4

Figure 1. Study area showing Colún and Chaihuín River and human settlements

Chaihuín and Cadillal Alto...... 103

xiii

Figure 2. Frequency of dog owners reporting American mink or Southern river otter

near their households and whether they detected a dog-mustelid interaction...... 104

Figure 3. Frequency of occurrence of river otter, mink and domestic dogs at camera

traps by otter latrine and random sites...... 105

Figure 4. Co-occurrence of dogs and American mink at an otter latrine site...... 106

Figure 5. Co-occurrence of mink and river otters at otter latrine sites...... 107

Figure 6. Schematic representation of the ‘bridge host’...... 108

Chapter 5

Figure 1. Schematic figure represent the conceptual idea of the ‘bridge host’...... 116

Figure 2. Schematic representation of transmission detections includind direct and

indirect interactions...... 118

xiv

CHAPTER 1

GENERAL INTRODUCTION

Infectious diseases are increasingly recognized as posing a risk to endangered carnivores

(Williams et al. 1988, Alexander and Appel 1994, Roelke-Parker et al. 1996). However, there is much that still needs to be learned about how disease is maintained and transmitted in carnivore assemblages. For example, there is increasing evidence that canine distemper virus, an important disease of concern in carnivores, can only be maintained/persist in a multi-host system in wild carnivores (Craft et al. 2008, Almberg et al. 2010) and that hosts can include domestic dogs and wild carnivore populations

(Cleaveland et al. 2000). What is not well understood is what drives the interactions that are needed to maintain disease in such a mixed species community. Canine distemper virus is generally modeled as being transmitted through direct contact. However, carnivore species do not normally interact directly except through conflict (one killing another) (Palomares and Caro 1999, Donadio and Buskirk 2006). How then, is distemper being maintained in these mixed species groups? How is disease being transmitted?

Most research on wildlife interactions have focused on intra-species interactions and behavior rather than inter-species interactions. Carnivore species are known to interact indirectly through the sharing of carcasses (Cooper 1991, Paquet 1992, Atwood 2006); however, other interactions must occur as species move on the landscape in overlapping home ranges (Ben-David et al. 1996). More work is needed to understand how and why

1 these interactions occur. This dissertation aims at elucidating some of the drivers of multi-species interaction that lead to infectious disease, and, particularly, CDV persistence and risk to endangered carnivores in Chile.

CANINE DISTEMPER VIRUS

In 1905, Henri Carré discovered the virus that causes canine distemper (Carre 1905).

However, the disease in dogs had been known since the mid-eighteenth century when outbreaks were reported following the importation of South American dogs into Europe

(Blancou 2004). These early studies of CDV focused on its deleterious effect in domestic dogs (Perdrau and Pugh 1930, Swango 1995) and resultant population impacts (Tennant et al. 1991). However in the last two decades there have been an increasing number of studies investigating CDV ecology and the impact of CDV on threatened wildlife populations (Williams et al. 1988, Alexander and Appel 1994, Alexander et al. 1996,

Roelke-Parker et al. 1996, Deem et al. 2000, Kapil and Yeary 2011, Sakai et al. 2013).

Canine distemper virus (CDV) belongs to the genus Morbillivirus from the family

Paramyxoviridae, and is a relative of a number of human and animal pathogens including human pathogen measles virus, rinderpest virus, phocine distemper virus (PDV), cetacean dolphin morbillivirus and feline morbillivirus (Domingo et al. 1990,

McCullough et al. 1991, Barrett et al. 1993, Gibbs et al. 2008, Woo et al. 2012). All

2 known Morbilliviruses have a highly specific host range except CDV which has a wide range of host species already identified, and more still being discovered, making its disease ecology highly complex.

Canine distemper virus is a single stranded RNA virus serologically classified as one serotype (Kapil and Yeary 2011), but the heterogeneity of amino acid variation in the H protein, a structural viral protein, allows classification by phylogenetic analyses (Greene and Vandevelde 2012). Today, 12 distinct CDV genotypes are classified, each with clear geographical distributions: American-1 and American-2 (in North America), Artic (in

Artic and Europe), Asia-1, Asia-2, Asia-3, European wildlife, South Africa, Argentina,

Rockborn-like (vaccine strain from Sweden), Mexican strain, and recently a new South

America-2 lineage (Calderon et al. 2007, Simon-Martinez et al. 2008, Woma et al. 2010,

Zhao et al. 2010, Kapil and Yeary 2011, Panzera et al. 2012).

The history of new CDV host discoveries during the last few decades has led to a better understanding of the multi-host nature of CDV ecology and epidemiology. An important breakpoint event occurs with the famous distemper outbreak in Serengeti lions in 1994 in

Tanzania (Roelke-Parker et al. 1996, Harder and Osterhaus 1997). At that time, cases of

CDV were considered rare within the felidae family. Since then, several species of big cats have been found to be susceptible including tigers, Asian lions, and leopards (Appel et al. 1994, Myers et al. 1997, Murray et al. 1999). Serological and/or molecular evidence of exposure has also been found in cougars ( Puma concolor ) and ocelots

(Leopardus pardalis ) in Brazil (Nava et al. 2008), Canada lynx ( Lynx canadensis ) and bobcats ( Lynx rufus ) in Canada (Daoust et al. 2009), and Iberian lynx ( Lynx pardinus ) in

Spain (Meli et al. 2010). This dramatically changed the scientific community’s understanding of CDV, which only a decade before had not considered it a disease of felids (Harder and Osterhaus 1997). At present, CDV host species are thought to include all members of the Carnivora order (Deem et al. 2000). Although, generally considered to be transmitted by direct contact, CDV has a history of ‘jumping’ between geographically separated populations. In 1988 an outbreak was observed in the Russian

Baikal Lake that caused mass die offs of several thousand Lake Baikal seals ( Phoca siberica ) with clinical signs resembling distemper in dogs (Grachev et al. 1989). Genetic studies by Mamaev et al. (1995) concluded that the virus was more closely related to

CDV than PDV and was an identical genotype to a strain found in an outbreak in harbor seals ( Phoca vitulina ) along Northern European sea coasts (Osterhaus et al. 1990).

Mamaev et al. (1995), also found that the virus was most likely transmitted from terrestrial carnivores to the marine mammals. Therefore, not only was there transmission across taxonomically diverse species but also transmission between different habitats

(terrestrial-aquatic).

In addition to carnivores, CDV has now also been found in a number of non-carnivore species as well. Polymerase chain reaction studies have found CDV in collared peccaries

(Tayassu tajacu ) (Appel et al. 1991) and Asian marmots ( Marmota caudata ) (Origgi et al. 2013). Serological exposure has also been observed in wild boars ( Sus scrofa ) and

4 sika deers ( Cervus nippon ) in Japan (Kameo et al. 2012); and Oni et al. (2006) described low but positive CDV antibody titers in Asian Elephants ( Elephas maximus ) in Thailand.

These non-carnivore species might play a role in CDV transmission dynamics but to date no studies have covered this topic.

In 1989, Yoshikawa et al. (1989) described the first CDV cases in non-human primates in

Macaque monkeys ( Macaca fuscata ) in Japan. More recent large outbreaks in captive and free-ranging monkeys ( Macaca fuscata and Macaca mulata ) in Japan and China have also been found (Sun et al. 2010, Sakai et al. 2013). Bieringer et al. (2013) determined that CDV only requires one amino acid exchange (mutation) in the H-protein to adapt to human receptors based on studies in human epithelial cells. Based on this finding, the authors conclude “when the goal of measles eradication may be achieved, and when measles vaccination will be stopped, CDV might eventually cross the species barrier to humans and emerge as a new human pathogen” (Bieringer et al. 2013).

DISTEMPER TRANSMISSION AND PATHOGENESIS

Based on controlled transmission experiments done in the 1950s and 1960s, CDV is characterized as being transmitted by aerosol and droplet exposure (Appel and Gillespie

1972). Most studies about transmission were designed by exposing infected dogs or ferrets to susceptible individuals or by direct viral inoculation (e.g. aerosol nasal mucosa or intravenous) (Dunkin and Laidlaw 1926, Appel and Gillespie 1972). Shen and

Gorham (1978) observed a positive relationship between contact duration and viral transmission in ferrets. Further work by Shen and Gorham (1980) determined that virus survival in the environment was 48 hours at 25 °C and 14 days at 5 °C. Because of these studies, CDV has traditionally been characterized as spread by aerosol through direct contact and has been thought to have poor viral survival in the environment. However, viral particles have been found in all bodily excretions including saliva, feces and urine

(Appel and Gillespie 1972, Elia et al. 2006, Saito et al. 2006, Acton 2007).

Once in the oral pharynx, the virus invades the local lymphoid tissue where it rapidly replicates, causing cell death and leading to systemic lymphopenia and immunosuppression (von Messling et al. 2006). Depending on the host immune response, the individual can recover completely and become seropositive or the virus can continue to replicate and then spread throughout the body by means of the lymphoid tissue. Common targets include the respiratory, gastrointestinal, urogenital and nervous systems. The replication phase takes 7-10 days and is not associated with clinical signs

(incubation period). Once the organ systems are invaded, however, systemic disease emerges including conjunctivitis, coughing, pneumonia, vomiting, diarrhea, and neurologic signs associated with encephalitis (Appel 1987, Appel and Summers 1995,

Greene 2006). During the symptomatic period the individual is infectious and sheds the virus in all bodily excretions (Newbury et al. 2009). Mortality rates range from 25%-

75% and depend on factors such as immune response, viral strain, co-infections, species,

6 and age (Greene and Appel 2006). Animals that recover have lifelong immunity against subsequent infections and no evidence of carriers exists (Appel et al. 1982, Appel 1987).

CDV AS A RISK TO ENDANGERED POPULATIONS

Very small, isolated host populations (as with endangered species) are much more susceptible to catastrophic extinction due to stochastic variation in their environment than are larger more resilient populations. This makes them particularly prone to extinction due to disease. In cases where the host population for a given disease is small and widely dispersed and the pathogen is highly transmissible and lethal, on the other hand, disease ecology theory predicts further declines in host abundances, a decline in interactions and, ultimately, pathogen collapse (Lyles and Dobson 1993). As a result, rare wildlife that live at low population densities are not likely to maintain persistent levels of pathogens in their population so long as the population is isolated from pathogen influx (Hess 1996,

Lafferty and Gerber 2002). The situation becomes more complex when a pathogen can infect multiple species (multi-host pathogen). This can allow persistence of a pathogen in the community of hosts even when one individual host species nears extinction during a multi-host outbreak (McCallum and Dobson 1995, Dobson and Foufopoulos 2001, De

Castro and Bolker 2005). This, of course, further increases the risk of extinction from disease in rare and endangered wildlife populations.

Because of their high trophic level, carnivores naturally maintain low population densities under free-ranging conditions. When carnivores become endangered or threatened, these low densities, by definition, become critically low, making individuals difficult to study and interactions among individuals infrequent (Palomares and Caro

1999, Gittleman et al. 2001). Based on the discussion of small populations and disease, above, one might expect that rare carnivores would be both less likely to maintain pathogens in their population, but also be more susceptible to catastrophic extinction during an acute outbreak. However, due to the challenges of working with low-density and rare species there have been few successful studies that can provide insight into the ecology of in situ carnivore populations and disease. Today, most of the empirical knowledge of CDV dynamics in free ranging carnivores comes from Africa and North

America where long-term ecological studies are providing some insight into CDV ecology and pathogenesis in the wild.

An example of how a disease outbreak can act as a stochastic event leading to extinction of rare carnivores is the Black footed ferret. The Black-footed ferret ( Mustela nigripes ) was considered extinct for almost 16 years until a ranch dog brought a killed specimen to its owner in Meeteetse, Wyoming in 1981 (Clark and Campbell III 1981). This event resulted in the discovery of a small colony of about 100 animals, and the Black Footed

Ferret became the most threatened mammal in North America. After continuous population monitoring, the population grew to 129 individuals in 1984. However, in

1985 a combined outbreak of plague in the ferret’s primary prey, prairie dogs, and CDV

8 in ferrets (Williams et al. 1988) led to a precipitous decline in the population and the last

18 ferrets were brought into captivity between 1985 -1987 as a means to save the population from extinction. Sympatric carnivores sampled for distemper found virus in badgers ( Taxidea taxus ) and coyotes ( Canis latrans ) (Williams et al. 1988), possibly the source of infection to the ferrets. Fortunately, today the species has recovered and several populations have been reestablished through highly intensive captive breeding, management and reintroduction efforts (Jachowski et al. 2011). This first historical case of CDV nearly causing the extinction of an entire carnivore species was an important alert for the future of conservation managers and the impact of infectious diseases (May

1986).

A good example of how multi-host population dynamics can influence CDV outbreaks in wild carnivores exists in the Yellowstone National Park wolf population. When

Yellowstone was created in 1872, wolves were considered pests and legally unprotected even within the park limits. The last wolf was killed in 1943 and its extinction resulted in a number of ecological cascade effects including the overabundance of Elk and the resultant decrease in Aspen tree regeneration (Fortin et al. 2005). As a result, in 1995, 31 wolves were reintroduced to the park (Bangs and Fritts 1996), and, without human hunting pressure, the wolf population rapidly increased (Almberg et al. 2012). Since reintroduction, the Yellowstone wolf population has experienced repeated and cyclic

CDV outbreaks in 1999, 2002, 2005 and 2008 with additional apparent outbreaks leading to a decrease in pup survival in 2002 and 2005 (Almberg et al. 2009, Almberg et al.

2012). To better understand this cyclic pattern of outbreaks, Almberg et al. (2010) used a demographic-multispecies-spatial model to determine that the periodic distemper epidemics were explained by ongoing depletion of immune individuals through normal mortality and the concurrent increase in susceptible individuals through recruitment of pups into the population. Ultimately this leads over several years to a population with a large number of young naïve individuals and few older immune animals creating adequate conditions for a new epidemic. Each outbreak then created a new population of animals with lifetime immunity to protect the pack from further outbreaks. However, the models also found that the number of potential hosts in this system at the scale of the study did not provide enough critical community size or the minimum number of susceptible hosts (Bartlett 1957) to allow pathogen persistence in this population. Only the periodic introduction or maintenance of the disease in an unknown carrier or alternative host species could explain their field data. This work implicates the multi- host nature of CDV as a critical factor in maintaining the disease in low-density carnivore populations. Domestic dogs are an unlikely reservoir in this case since distemper cases in veterinary clinics around Yellowstone are rare and dogs are, by and large, vaccinated for distemper in this region (Almberg et al. 2009, Almberg et al. 2010).

Nelson et al. (2012) studied serological exposure to CDV in wolves in Canada to determine if the seroprevalence was explained by proximity to human settlements (First

Nation Reserves) where unvaccinated domestic dogs could serve as a source of CDV.

These authors observed an inverse relationship from that expected with higher prevalence

10 in wolves far from human settlements, arguing for the possibility of other wild carnivores

(e.g. coyotes) being a more important co-host for pathogen persistence in wolves compared to domestic dogs. Thus, it is clear that domestic dogs are not the only source of disease transmission and outbreak in wild carnivore populations.

In most regions of Africa and South America, on the other hand, free-ranging and unvaccinated dog populations are common (Dalla Villa et al. 2010). Unfortunately, these domestic dog populations are often implicated as source reservoirs in CDV outbreaks in wildlife (Alexander and Appel 1994, Courtenay et al. 2001, Fiorello et al. 2006, Acosta-

Jamett et al. 2011). Dogs, unlike wild carnivores, tend to maintain artificially high population numbers and are increasingly encroaching on wild lands (Vanak and Gompper

2009, Young et al. 2011, Hughes and Macdonald 2013). The Serengeti ecosystem in

Africa is one of the best-studied sites for wild carnivore CDV ecology. Since 1960 the lions in this ecosystem have been continuously and intensively monitored in one of the longest continuous research study in any mammal, providing a unique foundation of information for understanding social, behavioral and ecological drivers of disease in a large social carnivore and top predator. In 1994, one third of the lions in Serengeti died

(~ 1,000 individuals) in an unprecedented CDV outbreak (Roelke-Parker et al. 1996).

Epidemiological studies determined the most probable source of infection was located in the abundant population of domestic dogs living in the surrounding local communities adjacent to the park (Cleaveland et al. 2000). However, although dogs were thought to be the source of CDV (Cleaveland et al. 2000), it was unclear how CDV reached the lions

11 from the domestic dogs. Dogs and lions do not commonly interact directly, so the classic transmission model of direct aerosol spread was unlikely. The 1994 lion CDV outbreak has provided some interesting modeling studies to answer this question. This outbreak affected 95% of the population and lasted for ~8 months (Craft et al. 2008). Models of the system determined that the observed outbreak pattern required more than one host

(lions), and that only the addition of hyenas and jackals in the model could explain the observed temporal and spatial pattern of spread among lion prides (Craft et al. 2008).

Later studies eliminated nomad lions as an important source of spread as well (Craft et al.

2009). Craft et al. (2009) also determined that the contact network structure within prides improved the model in relation to the observed disease pattern, suggesting that the lion pride structure facilitates the spread of distemper within lions. Finally, this study showed that the long duration of the outbreak was explained by a re-fuel of infected individuals from the co-hosts of CDV, hyenas or jackals (Craft et al. 2009). Jackals and hyenas, in turn, are more likely to surround villages and interact with domestic dogs, potentially leading to them acting as bridge host between the dogs and lions (Cleaveland et al. 2000,

Cleaveland et al. 2006). Interestingly, further studies in this system have determined that lions have both symptomatic and non-symptomatic outbreaks of distemper and that symptomatic outbreaks and mass-mortalities are associated with drought, increased predation on tick-infested buffalo ( Syncerus caffer ) and a resultant co-infection with the tick-borne parasite Babesia spp. (Munson et al. 2008). Clearly, these studies show that the transmission model for CDV and its impact in populations is not straight forward, and

12 that extensive information is needed as to what populations are interacting in these systems to drive persistence of multi-host pathogens in wild populations.

Thus, based on studies to date, we understand that multiple carnivore populations appear to be required to maintain CDV transmission and persistence in a wild carnivore assemblage in the cases described. However, whether the co-host of CDV is a domestic dog or a wild carnivore, the dynamics and drivers of such a multi-host system are little understood. How do such multi-host carnivore systems work? What drives interactions and disease transmission among carnivore populations? Free ranging carnivores are generally solitary and competitive. Interactions, if they exist, are usually antagonistic

(one carnivore killing another) or indirect (sharing a carcass for instance, or exposure to feces/urine). How does a disease generally considered to be passed by direct contact be maintained in such a system?

CARNIVORE INTERACTIONS

The behavioral heterogeneities among carnivores are enormous across families, species, within species (populations), within populations (male, females, ages) and even at the individual level (transients, social hierarchies) (Gittleman 2001). Most of their behavioral complexities are still under deep debate, but some insight is being gained as to how sympatric species interact. Interactions between carnivores are primarily mediated

13 by intra-guild competition, which occurs when a dominant species affects the fitness of the subordinate species resulting in reduced abundance, fecundity, survivorship or growth

(Begon et al. 1996). Competition may operate according to two processes. Exploitation competition occurs when one species depletes a limited resource and the loss of that resource negatively affects a subordinate species without need of direct interaction. To our knowledge no studies have demonstrated exploitation competition among carnivore species. Interference competition, in contrast, involves a direct interaction between species to the detriment of one (Begon et al. 1996) and includes intra-guild killing, food stealing or active avoidance (Palomares and Caro 1999, Creel et al. 2001, Caro and

Stoner 2003, Donadio and Buskirk 2006).

Clearly, these kinds of competition relationships might influence contact rates among carnivores, and, therefore, pathogen transmission. For example, pathogens transmitted by direct contact would be favored in cases where food stealing or killing occurs among species, resulting in a possible transmission event (i.e . hyenas-lions and coyotes – wolves). In cases where active evasion is the result of competition (i.e. reducing space or temporal overlap), environmental or vector borne pathogens may be transmitted between species (i.e. CPV, Toxoplasma gondii ) while direct transmission of pathogens will be unlikely.

Competition within carnivore species may also play a role in disease transmission risk.

For example, males or packs often are highly territorial and defend territories

14 aggressively against others of their own species (Grinnell et al. 1995, Rich et al. 2012).

This leads to a bi-phasic transmission model where disease is transmitted rapidly within packs, but lower transmission rates exist between packs (Haydon et al. 2002, Craft et al.

2008, Almberg et al. 2010). One form of intra species communication that is particularly prominent in the carnivores is scent marking (Kleiman 1966, MacDonald 1980). Some species even repeatedly deposit feces, urine or anal gland secretion at particular sites to mark territory (latrines) (Gorman and Trowbridge 1989, Hutchings and White 2000).

Such territorial marking behavior has a variety of purposes in carnivores: territory maintenance and defense, mediating encounters (i.e . reproductive status) and conveying information about availability/use of resources (Hutchings and White 2000). However, scent-marking behaviors could also potentially facilitate indirect disease transmission through fecal-oral or urine-transmission routes. This is particularly true in the latrine sites where multiple individuals are frequently and repeatedly depositing potentially contaminated biological material in a highly restricted site. These interactions are generally thought to be between individuals within a species, but could these behaviors also lead to transmission risk among species? These collateral disease transmission effects of carnivore behavior are poorly understood and bear further study (Page et al.

1999). How can we translate this ecological information on how and why carnivores interact on a landscape to allow multi-host diseases like CDV to persist despite low densities of individual species?

FRAMING THE RESEARCH PROBLEM

Studies of threatened, endangered, or reintroduced carnivore species in North America and in some parts of Africa, have supplied important information on the epidemiology of

CDV in these ecosystems over recent decades. What we understand, though, is what is happening (multiple carnivores are interacting to allow disease transmission and persistence in a multi-host system), but not why or how it is happening. This is particularly true for carnivores in forest ecosystems where capturing and visualizing rare carnivores can be challenging, especially sick individuals. Yet, the knowledge of the drivers of individual or interspecies interactions is crucial for the understanding of pathogen transmission ecology, especially in multi-host pathogen systems (Craft et al.

2011, Dietz et al. 2013). Therefore, it is the goal of this dissertation to focus on understanding the causes and description of interactions between sympatric carnivore species in order to better elucidate and manage disease transmission risk to endangered carnivores. Specifically, this work used a multi-carnivore system composed of domestic dogs, invasive American mink and endangered southern river otters to investigate drivers of carnivore species interactions in the Southern Temperate Forest of Chile. The first study (Chapter 2) investigated how differences in human use and management of dog populations changed interaction risk between domestic dogs and wild species. Chapter 3 investigated movements of rural dogs and how landscape shapes dog foray behavior into protected areas and wildlife habitat. Chapter 4 elucidates how an invasive species, the

American mink, is driving changes in a contact framework of domestic dogs and river

16 otters and whether the introduction of this species is increasing the risk of inter-species

CDV transmission in this assemblage. Finally, Chapter 5 (Conclusions) focuses on how these new findings will inform management and lead to future research focused on the importance of host behavior and species interaction in disease ecology and conservation.

CHAPTER 2

DOMESTIC DOGS IN RURAL COMMUNITIES AROUND PROTECTED

AREAS: CONSERVATION PROBLEM OR CONFLICT SOLUTION?

INTRODUCTION

Dogs were domesticated over 14,000 years ago (Vila et al. 1999), and as human populations have grown, so have populations of dogs. Today, the world dog population is estimated at 700 million (Hughes and Macdonald 2013). In developing countries, rural dogs, defined by Vanak and Gompper (2009) as free-ranging dogs associated with humans and pastoral or farm activities, are of particular concern as human populations expand into borderland and natural areas in search of affordable and productive farm land

(Wittemyer et al. 2008, Vanak and Gompper 2009, Young et al. 2011). Recent studies have found that rural dogs, particularly free-ranging dogs, can threaten wildlife through predation, competition or as sources of infectious diseases (Young et al. 2011, Hughes and Macdonald 2013). At the same time, dogs are being promoted as a nonlethal strategy for abating human-wildlife conflict in some areas (Rigg 2001, Gehring et al. 2010b,

González et al. 2012). For example, guarding dogs can protect livestock from carnivore predation, thus reducing human-wildlife conflict and protecting vulnerable carnivore populations from retaliation killing (Andelt 1992, Andelt and Hopper 2000, Gehring et al.

2010b). While these studies typically have focused on the interaction between domestic

18 dogs and wildlife, to date little research has been conducted regarding the reasons rural families keep dogs and how these roles impact dog-wildlife interactions. A specific understanding of the motivation for dog ownership in rural communities and how dog populations are managed is lacking, increasing the likelihood of detrimental or unpopular dog management policies and strategies around conservation areas.

The specific objective of this study was to test whether the role and management of dogs influences the nature and frequency of dog-wildlife interactions. Using a survey of households in rural communities in Chile, we investigated the relationship between farm management practices and farmer-reported dog-wildlife interactions. In addition, the survey provided important information on dog management and demography, both relevant for improving dog population management in the future, particularly in wildlife conservation areas. By examining the role of dogs from a subsistence farming perspective, this study provides key insights into the role of the dog in farming and conservation. This study elucidates the critical but socially sensitive tradeoff between dog as farm-protector and dog as wildlife threat.

METHODS

Study area

The study area was located in the coastal range of the Chilean Temperate Forest between latitude 39º41' and 39º58'S. Between 11 and 76 households were interviewed in 5 rural communities (Curiñanco, Huape, Chaihuín, Huiro and Cadillal Alto) adjacent to three different protected areas (Curiñanco Reserve, Alerce Costero National Park and the

Valdivian Coastal Reserve [VCR]) (see Fig. 1). The economy of these rural communities is based mainly on extraction of natural resources from the sea (seafood, algae), livestock production (primarily cattle, sheep and poultry) and forest resource utilization (timber and native berries). In all five rural communities more than 75% of households were classified as below the poverty level (monthly income < 60 US$/household) (Delgado

2010). Land cover in these areas is primarily characterized as native evergreen forest located within protected areas in the higher elevations, grazing pastures in the lowlands and outside of the protected areas, and exotic plantations (eucalyptus and pines) mostly outside of the protected areas (except in the VCR where eucalyptus plantations are also present). Average annual rainfall is 2.5 m per year and annual average temperature is

12°C (average minimum of the coldest month of the year: 5°C; average maximum of the hottest month of the year: 17°C) (Luebert and Pliscoff 2005). All of the study sites were within the Valdivian Ecoregion, an area that has been labeled a global biodiversity hotspot (Olson and Dinerstein 1998, Brooks et al. 2006).

Free-ranging dogs are common in the study area, and both predation on the Southern pudu ( Pudu puda ) and European hare ( Lepus europaeus ) and interference competition with the chilla fox ( Lycalopex griseus ) have been reported (Silva-Rodriguez et al. 2010,

Silva-Rodriguez and Sieving 2011). Other wild mammal species (>1 kg) observed in the study site and potentially affected by the presence of dogs are the guigna ( Leopardus guigna ), puma ( Puma concolor ), Molina’s hog-nosed skunk ( Conepatus chinga ), lesser grison ( Galictis cuja ), Southern river otter ( Lontra provocax ), marine otter ( Lontra felina ), coypu ( Myocastor coypus ) and the invasive American mink ( Neovison vison ).

Species of conservation concern are the Southern river otter (endangered), marine otter

(endangered), guigna (vulnerable) and the Southern pudu (vulnerable) (IUCN 2013).

Survey Instrument and Implementation

From October to December 2010 we attempted to interview one adult per household in the five rural communities (Fig. 1). The total number of households estimated by local park rangers was 245 households. In instances where we did not find anyone at the house during a scheduled visit, we conducted up to three repeat visits. Information collected included the number of dogs per household, dog management practices including the type of food provided and whether there was any movement restriction of the dogs (i.e. completely free-roaming or restricted to some degree), dog demographics (age, sex, reproduction, mortality and source of dogs), presence of farm animals (poultry, cattle,

21 sheep), observed dog attacks/harassment of wildlife, and measures of farm animal protection, including dogs. The questionnaire was comprised of 15 questions and was based on instruments used for studies with similar objectives (Fiorello et al. 2006, Silva-

Rodriguez and Sieving 2011) (see Supplementary Material). It was not expected that farmers would avoid questions related to dog-wildlife interactions since previous work in the region found that residents openly reported wildlife interactions including killing threatened species (Stowhas 2012). Oral informed consent was provided by an adult in each interviewed household in a standardized format. The general purpose of the survey was explained in Spanish to each interviewee and consent was obtained by having the interviewee state that they agreed to participate. The data was analyzed anonymously.

The use of this survey and approach was approved by the Institutional Review Board at the University of Minnesota (1304E31141).

Dog utility and wildlife interactions

One way in which the questionnaires were used was to determine if owned rural dogs around the protected areas were used to protect poultry and sheep against wild predators.

We quantified the wildlife species and domestic dog/wildlife interactions each respondent observed during the prior year. A dog-wildlife interaction was defined as a killing or harassment event by the dog to either: a) other conflicting carnivores (including mink, fox, guigna and puma), referred to as dog-carnivore interaction, or b) herbivores or prey

22 animals, particularly European hare and Southern pudus, referred to as dog-prey interaction. To assess the potential influence of the dog owner in the dog-carnivore interaction, we determined the percentage of dog owners who reported chasing off a particular predator species using dogs.

Statistical analysis

We used generalized linear models with a logit link function to assess the probability of dog-carnivore and dog-prey interactions (Agresti 2002). A binary response variable was based on the dog owner reporting the presence/absence of a dog interaction with wildlife during the last year. Two different outcomes were modeled for wildlife, one for dog-wild carnivore interactions (dog-carnivore, hereafter) and another one for dog- prey interactions. For both models, we used the following predictor variables: 1) number of dogs per house (DOGS), 2) presence of chickens (POULTRY); 3) presence of sheep

(SHEEP); 4) village (LOCATION TYPE), and 5) whether commercial dog food was included in the diet of the household dogs (FOOD).

To determine significant predictors for dog-carnivore and dog-prey interactions we used an information-theoretic approach. A priori models that best explained each type of interaction were developed by creating all additive model combinations of the predictors named. We hypothesized that a higher number of dogs (DOGS) would increase the probability of dog-wildlife interaction at the household level. Location (LOCATION

TYPE) used as a covariate because of differing ecological conditions at each village ( i.e. habitat composition around localities). Specifically, the town of Curiñanco was used as the reference location type (dichotomous variable: Curiñanco versus the rest of the villages: Huape, Huiro, Cadillal Alto and Chaihuín) because this locality was closer to the main city of Valdivia (Fig. 1) and therefore possessed less forest coverage and more urban influence. This variation could potentially explain differences in human behaviors and carnivore composition and abundance. Chickens (POULTRY) are preyed upon by most small wild carnivores present in the study area (guigna, American mink, lesser grison and chilla fox), while sheep (SHEEP) are mostly preyed upon by the puma. Other farm animals such as cattle are infrequently attacked by predators (Stowhas 2012).

Finally, the use of commercial dog food (FOOD) was included because previous studies have demonstrated that the presence of poorly fed dogs correlates with events of dogs killing wild vertebrates (Silva-Rodriguez and Sieving 2011).

For model selection purposes we utilized the Akaike information criterion corrected for small sample size (AICc). We estimated Akaike differences ( AICc) and weights (wi ) to determine the relative support for a model given the data for the set of candidate models.

To account for model selection uncertainty we averaged the estimates of the coefficients of the main effect variables in all models. To determine the relative importance of the predictor variables we calculated the sum of the Akaike weights over all of the models in which the parameter of interest appeared (Burnham and Anderson. 2002). The magnitude of the effect of each predictor variable on the response variable was

24 determined by the odds ratio and 95% confidence interval of averaged estimates.

Goodness of fit for the global candidate models and best selected model for dog- carnivore and dog-prey interactions was assessed using the unweighted sum of squares test (Hosmer et al. 1997). All analyses were performed using the packages glm2, MuMIn and rms in R (R Development Core Team 2011)(R code provided in Supplementary

Material).

RESULTS

We interviewed 146 households available for inclusion in the study (59.6% of total households) with only two people refusing to accept the interview. Of these, 123 respondents had dogs and provided information about dog demography and management

(Table 1). We observed an adult-dominated age structure, highly skewed towards males, with a total population size of 218 among the interviewed households. During the year

2010 only 14 dogs (6.5% of total dogs) were added to the population through breeding

(recruitment) out of the 39 pups born. Some of the dog owners indicated that females in heat are problematic because they attract many male dogs, resulting in dogfights and negatively affecting the body condition of these working dogs.

In this study, we refer to dogs brought from other localities as ‘imported’ dogs in order to differentiate them from stray dogs that are adopted by the household. Importation of dogs ( n=27) from nearby towns or localities was a more important input to the population

25 than recruitment (12.3% versus 6.5%), and most of these imported dogs were males

(Table 1). Dog owners reported killing adult dogs that attacked domestic animals and unwanted pups from litters (Table 1). Among respondents, a significantly higher proportion (64%; P ≤ 0.005) were willing to spay female dogs compared to neutering male dogs (17.1%). Significantly more female dogs were neutered than males (Table 1).

In the previous year, 53 adult dogs died due to the following causes: illness (30.1%), car collisions (26.4%), age (20.7%), or killed by owner (7.5%). In addition, 3 dogs disappeared. The proportion of people that reported having killed a dog at least once in their lifetime was 27.1%, all related to attacking or killing farm animals. Nine households reported killing 14 different dogs (not owned by them) during the previous year because they had killed farm animals.

No feral dogs (wild dogs completely independent of human resources (Vanak and

Gompper 2009)) were reported by the respondents or observed by the research teams throughout the study period. However, most dogs in the survey were allowed to roam free (92.2%). Dog diet was based primarily on homemade food (73.2%), commercial food (52.0%) and wheat bran (40.5%). When asked about food provision for the dogs during the previous day, 6.5% of dog owners reported that no food had been provided.

Less than half the respondents vaccinated (33.6%) or dewormed (41.9%) their dogs.

Seventy-seven percent of interviewed households owned at least one kind of farm animal.

Poultry, cattle and sheep were the most common farm animals present (Fig. 2), and the

26 most common means of protecting these animals against predators were enclosures and use of dogs (Fig. 3).

Respondents reported having seen most of the listed wildlife species during the previous year. Exceptions include lesser grison and coypu that were observed by only two and nine respondents, respectively. Dog-wildlife interactions during the previous year were reported for all species (Table 2) with the exception of the grison and river otter where no interactions were reported. Most interactions were detected near the household. During the study period, dog owners were observed rescuing two Southern pudus from dog attacks and bringing them to the VCR facilities for veterinary support. During informal conversations with four different farmers, it was consistently expressed that the farmers do not like dogs that grow accustomed to making long-distance hunting forays because they believe that this behavior negatively affects the dog’s body condition, and if repeated frequently will reduce the dog’s utility in guarding the household and the farm animals.

The variables included in the best-ranked models ( ∆AICc < 2) to explain the odds of dog- carnivore interactions included POULTRY and DOGS. Averaged parameter estimate of presence of poultry in the household was positively associated with dog harassment and killing of carnivores (see table S1 for all competing models). A marginally positive association with the probability of a dog-carnivore interaction was also attributable to the number of dogs (DOGS; Fig. 4). Adequate food (FOOD), locality (LOCATION TYPE)

27 and sheep ownership (SHEEP) were poor predictors for dog harassment of carnivores

(Fig. 4). For the dog-prey interactions, the best candidate models ( ∆AICc < 2) included all of the predictors. Averaged parameter estimates for dog-prey interactions were positively associated with DOGS and negatively associated with FOOD (Fig. 4 and Table

3). The remaining averaged predictor parameters were not considered statistically significant.

The percentages of dog-carnivore interactions that were motivated by dog owners actively encouraging the dog to chase off a predator species were 75%, 62.5%, 38.9% and 25% for American mink, puma, Chilla fox and guigna, respectively. The only dog- prey interaction that was motivated by humans was the European hare, where 10% of the total interactions were encouraged by dog owners.

DISCUSSION

This study investigated dog – wildlife interactions in the context of farmer management and dog utility. In this study rural dogs were predominantly free-roaming and their interactions with wildlife were significantly influenced by their role on the farm as well as by their management. Dogs appeared to interact mostly with carnivores at the encouragement of their owners in the context of protecting livestock. In contrast, dog interactions with prey species were mostly undesirable from the farmer’s perspective and

28 seemed to be driven more by hunger and inadequate diet. The skewed dog sex ratio and high rate of externally-acquired dogs showed that the dog population was heavily managed by farmers, indicating a potential opportunity for informed interventions aimed at minimizing interactions with wildlife.

Davlin and VonVille (2012) found that rural areas in developing countries have more houses with dogs and more free-roaming dogs than urban areas, a situation similarly observed in the rural dog population studied here. In this study, people strongly selected males over females thereby generating a strong bias in terms of sex ratios and, as a consequence, a low reproductive rate and poor recruitment of new animals to the population through breeding. This skewed sex ratio is consistent with other studies of domestic dogs in developing countries with increasing male bias in rural areas such as Sri

Lanka (Matter et al. 2000), Mexico (Flores-Ibarra and Estrella-Valenzuela 2004), and

Madagascar (Ratsitorahina et al. 2009). A previous study in northern Chile also found a male predominance of 56% in cities, 74% in towns and 83% in rural areas (Acosta-Jamett et al. (2010). Matter (1989) found that male bias in dogs in Turkey and Tunisia could be attributed to population control through selective culling and abandonment of female pups. The skewed sex ratio in the present study was influenced by culling as well as the acquisition of predominantly male dogs from other areas into this population, female dogs being considered difficult to manage and detrimental to the work capacity of the males during periods of estrus. A third, but minor means of control in this population was surgical sterilization. When used, it was targeted at females, possibly to minimize

29 management problems associated with estrus. Even though, in general, there are marked behavioral differences between male and female dogs (Hart 1995), there is some evidence that working dog performance is not affected by the sex of the dog (Svartberg

2002). In support of this, farmers stated motivation to avoid female dogs was associated with their reproductive behavior (estrus and pregnancy), and not their performance.

Given the management of the dogs and the low number of females in this study area, neutering programs would likely have little effect on the already human-controlled dog population but might improve dog welfare to address dog killings by humans (unwanted bitches) if accompanied by educational programs.

In our study we found that the interactions between dogs and wild animals were associated with their management and care. Interestingly, the models that best explained dog attacks on carnivores included poultry and the number of dogs, whereas the models that best explained interactions with prey included type of food provided and the number of dogs as the strongest predictors. Having a greater number of dogs seemed to be associated with a higher frequency of dog attacks on wild animals. This may be related to sampling effects (i.e., the probability of observing at least one dog attacking wildlife is higher for owners of multiple dogs) or to pack hunting behaviors. Silva-Rodriguez and

Sieving (2011) reported that dogs fed commercial food attacked prey species less often than those fed on low-quality food such as wheat bran or household leftovers. While our study detected a similar association for dog-prey species interactions, we also showed that dog-carnivore interactions were better predicted by livestock ownership than food

30 provision.

The fact that dog-prey and dog-carnivore interactions were predicted by different variables has significant implications for wildlife conservation. Minimizing impacts on prey will require different dog management strategies compared to impacts on carnivores in communities surrounding wildlife sanctuaries. In terms of dog-carnivore interactions, dogs that belonged to households with livestock had a higher likelihood of interacting with carnivores. In these circumstances, the use of dogs selected specifically for certain behavioral traits (i.e . shepherd or guarding breed dogs) could potentially minimize wildlife-human conflict and revenge killing of wildlife. Domestic dogs have been used to guard livestock against predator attacks throughout recorded human history (Rigg 2001) and continue to be used in this manner worldwide (Rigg 2001, Gehring et al. 2010b,

González et al. 2012). In a study by González et al. (2012) that was conducted in the northern Patagonian steppe of Argentina, 37 puppies of mixed-breed guarding dogs were delivered to 25 herders. These dogs were effective at reducing both goat losses and retaliatory killing. In the current study, dogs were used to protect poultry and other livestock, and dog-carnivore interactions were directly related to this key function on the farm. In fact, dog owners encouraged their dogs to chase carnivores seen as threats to their livestock. The fact that the presence of dogs is perceived by farmers to reduce carnivore-caused livestock losses may also result in a reduced retaliatory mortality of carnivores, suggesting that dogs whose primary function is to guard livestock indirectly contribute to reducing human-caused mortality of carnivores (Gehring et al. 2010b,

González et al. 2012).

The situation with dog-carnivore interactions is in direct contrast to the situation with dog-prey interactions as identified in this study. Here, dog-prey interactions are driven primarily by hunger and, for the most part, are perceived by the dog owner as negative and distracting the dog from the primary function of livestock and homestead protector.

These behaviors are clearly detrimental to wildlife populations. For example, a study in

New Zealand documented a single German shepherd that was estimated to have killed approximately 500 kiwis ( Apterix australis ) out of a population of 900 within a six-week period (Taborsky 1988). Recent work by Silva-Rodríguez and Sieving (2012) using camera traps in the same study area as our current study determined that dogs shaped the distribution of pudu by lethal (predation) and non-lethal (avoidance) means. In this study, dogs occasionally harassed or killed pudus despite the fact that people held highly positive attitudes toward pudus (Stowhas 2012) and actively tried to save pudus from dog attacks. In addition, some interviewees reported that their dogs killed domestic livestock, which is a frequent problem in southern Chile (INE 2011) and elsewhere (Schaefer et al.

1981, Ciucci and Boitani 1998, Sundqvist et al. 2008), further expanding the potential unintended negative consequences of domestic carnivores. Thus, although dogs can play an important role in protecting livestock and farms from wildlife pests, and thereby minimize human-wildlife conflict and retaliation, dogs also have direct impacts on wildlife and livestock populations through predation. This hunting activity is generally considered undesirable by dog owners and is an activity that can be minimized through a

32 change in diet and management. This provides an interesting opportunity to manage dog populations in a way that maximizes wildlife conservation while also protecting the rural farmers’ way of life and livelihoods. The fact that most dogs added to this rural population were obtained from localities outside of the study area is important information for future management interventions. Because dogs are imported anyway, interventions could focus on improving the health and behavior of these imported animals. Interventions, for example, could include the provision of guard dog breeds allowing a potential increase in the effectiveness of the guarding behavior of the dog while minimizing the predatory behavior. This increased effectiveness might also result in a reduction in the number of dogs in the area (González et al. 2012). Imported dogs could also be vaccinated for diseases of concern to wildlife prior to importation to minimize risk of disease transfer to endangered wildlife. In addition, governmental or conservation agencies could also include support for dog sterilization and educational campaigns around providing adequate food provision for dogs, health care, neutering and management of problematic dogs (e.g . dogs that harass pudus or livestock) at the same time that new dogs are provided.

Clearly, as dog populations grow, it will be critical to improve dog management close to sensitive conservation areas. At the same time, dogs clearly play a critical role as working members of households (e.g. herding, guarding, and hunting) in rural communities in developing countries. Thus, improving management of problematic behaviors in rural dogs around protected or natural areas will require a balanced approach

33 that integrates the significant benefit that dogs provide for rural residents into the management strategy. By engaging rural communities directly, this study shows that dog roles and management in the household are significant factors in dog-wildlife conflict.

By understanding these drivers of interaction, interventions can be targeted to optimize beneficial behaviors for the farmer while minimizing detrimental wildlife interactions.

This provides an important model for future management of anthropogenic pressure on conservation areas.

Tables & Figures

Table 1. Demography and management of domestic dogs in rural areas around three protected areas in Southern Chile.

Demographic parameters Responses Number of households interviewed 146 Households with dog ownership (%) 123 (84.2) Total dogs 218 Average number of dogs per household (SE) 1.7 (0.06) Average dog age (years old) (SE) 4.5 (0.02) Male: female ratio 7.3:1 Number of litters in prior year 9 Number of total pups 39 Pup survival rate before 2 months old 0.36 a Number of dogs recruited from breeding 14 Percentage of male dogs castrated 1.6 Percentage of female dogs neutered 26.6 Percentage of dogs obtained locally 44.9 Number of imported dogs less than 1 year old 27 b Number of adult dog mortalities 53 c

Household management Percentage of dogs allowed to roam free 92.2 Percentage of dogs fed commercial food 52.2 Percentage of dogs used for farm animal 69.7 protection a 36% of them killed by the dog owner; b 92.6% males and 7.4% females; c 7.5% killed by the dog owner.

Table 2. Number of households with dogs (n = 123) indicating dog-wildlife interactions observed during the previous year.

Number households with dog Species interactions (%) Poultry predators Fox 22 (17.9) Guigna 10 (8.1) American mink 6 (4.9) Lesser grison 0 Livestock predator Puma 13 (10.6) Other carnivores River otter 0 Marine otter 1 (0.8) Molina's skunk 8 (6.5) Herbivores European hare 73 (59.3) Southern pudu 6 (4.9) Coypu 1 (0.8)

Table 3. Summary of model selection to estimate the probability of owned dog interactions with carnivores and prey species. Models are ranked by AICc values. Columns include the number of variables (K), Akaike’s Information Criterion (AICc), distance from the lowest AICc ( AICc), and Akaike’s model weight ( ωi). Models showed include only those with a ∆AICc ≤ 2. See table1 in Supplementary material for all competing models.

Competing models K AICc ∆AICc ωi Dog-carnivore interactions POULTRY 2 144.45 0.00 0.21 DOGS+POULTRY 3 146.03 1.58 0.09

Dog-prey interactions DOGS+ FOOD 3 158.25 0.00 0.27 DOGS+POULTRY+ FOOD 4 159.25 1.00 0.16 POULTRY+DOGS+SHEEP 4 159.92 1.67 0.12 DOGS+LOCATION TYPE+ 4 160.24 1.99 0.10 FOOD

Figure 1. Map of the study area in the Coastal Range of southern Chile.

Figure 2. Farm animal ownership in rural households around three protected areas in southern Chile.

Figure 3. Measures of farm animal protection against carnivore predators reported by rural interviewees around three protected areas in southern Chile.

Figure 4. Model-averaged parameter estimates shown as odds ratios and 95% confidence intervals for explaining rural dog interactions with carnivore and prey species. Odds Ratios < 1 indicate a negative association with occurrence while Odds Ratios > 1indicate a positive association with occurrence.

SUPPLEMENTARY MATERIAL

Dogs owner questionnaire

1. ¿Cuantos perros tiene en este momento? (hembras, machos). (At present, how many dogs do you have? males and females).

2. ¿Qué edad tienen?. ( How old are they?).

3. ¿Cual es el origen de sus perros?. (Where did you get them?).

4. ¿Como maneja a sus perros? permanentemente libre, amarrado, libre solo de día. (How do you manage your dogs? Continuously free, leashed, leashed only during the day ).

5. ¿Están vacunados o desparasitados?. ( Are they vaccinated or treated against parasites? ).

6. Durante el último año ¿cuantos perros ha tenido?. (During the last year, how many dogs/cats have you had? ).

7. Durante el año pasado, ¿cuantos perros se han muerto? edad y causa. (During the last year, how many of your dogs have died?. Age and cause of death ).

8. ¿A matado alguna vez un perro? ¿Fecha de la ultima vez? ¿Cual fue el motivo? (Have you ever killed a dog? Date to the last time? What was the reason? ).

9. ¿Cuantas veces han parido las hembras en el último año? ¿Cuando?. ( During the last year, how many of the females gave birth? When? ).

10. ¿Cuantos cachorros en cada parto? ¿qué hizo con ellos?. ( How many puppies were born per litter? what did you do with them? ).

11. ¿Cual es el principal alimento de sus perros? ¿Cual fue su comida ayer? (What is the main source of food of your dogs? What was their meal yesterday?)

12. ¿Ha visto alguna vez alguno de estos animales? (Se mostraron 2 fotos para identificacion de cada especie silvestre) ¿donde? ¿cuando? ¿que hacían? ¿qué

hizo usted?. ( Have you ever seen one of these named species? [Two pictures for each wild species were provided for identification] Where? When? What did they do? What did you do? ).

13. ¿Ha cazado o atacado su perro alguna de las siguientes especies en el último año? cuantos, donde, y cuando: Pudu, zorro, guiña, vison, chingue, huillin, coipo, liebre. (During the last year, has your dog killed or attacked any of the following species? how many, where and when: Pudu deer, foxes, lesser grisson, mink, guigna, skunk, river otter, marine otter, coypu. ).

14. ¿Cuantos de los siguientes animales tiene (gallinas, ovejas, vacas, cerdos, cabras, otros)?. (How many of the following animals do you have: hens, sheep, cows, pigs, goats, others .)

15. ¿Qué medidas de protección toma contra predadores de sus animales domésticos? Armas, perros, veneno, encierro, trampas, otras, nada. ( What measures do you use to protect your farm animals against predators?: shotguns, dogs, poison, closure, traps, others, nothing. )

Interviewed person: Sex: Female Male

Age (years):

Occupation:

Locality:

Number of people habiting the house: Adults ___ Children ___

Table S1. Summary of model selection to estimate the probability of owned dog interactions with carnivores and prey species. Models are ranked by AICc values. Columns include the number of variables ( K), Akaike’s Information Criterion (AICc), distance from the lowest AICc ( AICc), and Akaike’s model weight ( ωi).

Competing models K AICc ∆AICc ωi Dog-carnivore poultry 2 144.45 0.00 0.21 interactions dogs+poultry 3 146.03 1.58 0.09 poultry+location type 3 146.46 2.01 0.08 poultry+food 3 146.48 2.03 0.08 poultry+sheep 3 146.55 2.10 0.07 intercept 1 147.06 2.61 0.06 dogs 2 147.94 3.49 0.04 dogs+poultry+location type 4 148.03 3.58 0.03 dogs+poultry+ food 4 148.08 3.63 0.03 poultry+dogs+sheep 4 148.16 3.71 0.03 sheep 2 148.45 4.00 0.03 poultry+location type+ food 4 148.48 4.02 0.03 poultry+location type+sheep 4 148.60 4.14 0.03 poultry+ food +sheep 4 148.61 4.16 0.03 food 2 149.12 4.67 0.02 location type 2 149.12 4.67 0.02 dogs+sheep 3 149.60 5.14 0.02 dogs+location type 3 150.01 5.55 0.01 dogs+ food 3 150.03 5.58 0.01 dogs+poultry+location type+sheep 5 150.21 5.75 0.01 dogs+poultry+ food +sheep 5 150.25 5.80 0.01 location type+sheep 3 150.48 6.03 0.01 food +sheep 3 150.54 6.09 0.01 poultry+location type+ food +sheep 5 150.65 6.20 0.01 location type+ food 3 151.21 6.76 0.01 dogs+location type+sheep 4 151.62 7.16 0.01 dogs+ food +sheep 4 151.72 7.26 0.01 dogs+location type+ food 4 152.12 7.67 0.00 dogs+poultry+location type+ food +sheep 6 152.28 7.83 0.00 location type+ food +sheep 4 152.59 8.14 0.00 dogs+location type+ food +sheep 5 153.76 9.30 0.00

Dog-prey dogs+ food 3 158.25 0.00 0.27 interactions dogs+poultry+ food 4 159.25 1.00 0.16 poultry+dogs+sheep 4 159.92 1.67 0.12 dogs+location type+ food 4 160.24 1.99 0.10 dogs+poultry+ food +sheep 5 161.33 3.08 0.06 dogs+location type+ food +sheep 5 162.03 3.78 0.04 dogs 2 162.44 4.19 0.03 poultry+ food 3 162.88 4.63 0.03 food 2 163.02 4.77 0.02 dogs+poultry 3 163.03 4.77 0.02 dogs+poultry+location type+ food +sheep 6 163.50 5.25 0.02 dogs+sheep 3 164.03 5.78 0.02 food +sheep 3 164.03 5.78 0.02 dogs+location type 3 164.55 6.29 0.01 location type+ food 3 164.76 6.51 0.01 poultry+ food +sheep 4 164.78 6.53 0.01 poultry+location type+ food 4 164.84 6.59 0.01 poultry+dogs+sheep 4 165.10 6.84 0.01 dogs+poultry+location type 4 165.15 6.90 0.01 location type+ food +sheep 4 166.02 7.77 0.01 dogs+location type+sheep 4 166.14 7.89 0.01 poultry 2 166.74 8.49 0.00 poultry+location type+ food +sheep 5 166.84 8.59 0.00 dogs+poultry+location type+sheep 5 167.24 8.99 0.00 intercept 1 167.43 9.18 0.00 sheep 2 168.36 10.11 0.00 poultry+sheep 3 168.64 10.39 0.00 poultry+location type 3 168.84 10.59 0.00 location type 2 169.47 11.22 0.00 location type+sheep 3 170.45 12.20 0.00 poultry+location type+sheep 4 170.77 12.52 0.00

R code to model dog-carnivore and dog-prey interactions: require("glm2") require(MuMIn) require(AICcmodavg) require("MASS") require("psych") require("glmulti")

Datos #data set with responde and predictor variables

##Model building of different candidate models to model the probability of dog- carnivore interactions: NASPG <- glm2(Car ~ Ndog+Hens+Location+Pellet+Sheep, data=Datos, binomial) NASG <- glm2(Car ~ Ndog+Hens+Location+Sheep, data=Datos, binomial) NAPG <- glm2(Car ~ Ndog+Hens+Pellet+Sheep, data=Datos, binomial) NSPG <- glm2(Car ~ Ndog+Location+Pellet+Sheep, data=Datos, binomial) ASPG <- glm2(Car ~ Hens+Location+Pellet+Sheep, data=Datos, binomial) NPG <- glm2(Car ~ Ndog+Pellet+Sheep, data=Datos, binomial) NSG <- glm2(Car~ Ndog+Location+Sheep, data=Datos, binomial) SPG <- glm2(Car ~ Location+Pellet+Sheep, data=Datos, binomial) APG <- glm2(Car ~ Hens+Pellet+Sheep, data=Datos, binomial) ASG <- glm2(Car~ Hens+Location+Sheep, data=Datos, binomial) ANG <- glm2(Car ~ Ndog+Hens+Sheep, data=Datos, binomial) NAS<- glm2(Car ~ Ndog+Hens+Location, data=Datos, binomial) ASP <- glm2(Car ~ Hens+Location+Pellet, data=Datos, binomial) NAP <- glm2(Car ~ Ndog+Hens+Pellet, data=Datos, binomial) NSP <- glm2(Car ~ Ndog+Location+Pellet, data=Datos, binomial) PG <- glm2(Car ~ Pellet+Sheep, data=Datos, binomial) NG <- glm2(Car ~ Ndog+Sheep, data=Datos, binomial) NA1 <- glm2(Car ~ Ndog+Hens, data=Datos, binomial) SP <- glm2(Car ~ Location+Pellet, data=Datos, binomial) AS <- glm2(Car ~ Hens+Location, data=Datos, binomial) SG <- glm2(Car~ Location+Sheep, data=Datos, binomial) AG <- glm2(Car ~ Hens+Sheep, data=Datos, binomial) AP <- glm2(Car ~ Hens+Pellet, data=Datos, binomial) NP <- glm2(Car ~ Ndog+Pellet, data=Datos, binomial) NS <- glm2(Car ~ Ndog+Location, data=Datos, binomial) N <- glm2(Car ~ Ndog, data=Datos, binomial) A <- glm2(Car ~ Hens, data=Datos, binomial) P <- glm2(Car ~ Pellet, data=Datos, binomial) S <- glm2(Car ~ Location, data=Datos, binomial) G <- glm2(Car ~ Sheep, data=Datos, binomial) Int<- glm2(Car ~ 1, data=Datos, binomial)

ScarMod

##Model building of different candidate models to model the probability of dog-prey interactions:

NASPG <- glm2(Preys ~ Ndog+Hens+Location+Food+Sheep, data=Datos, binomial) NASG <- glm2(Preys ~ Ndog+Hens+Location+Sheep, data=Datos, binomial) NAPG <- glm2(Preys ~ Ndog+Hens+Food+Sheep, data=Datos, binomial) NSPG <- glm2(Preys ~ Ndog+Location+Food+Sheep, data=Datos, binomial) ASPG <- glm2(Preys ~ Hens+Location+Food+Sheep, data=Datos, binomial) NPG <- glm2(Preys ~ Ndog+Food+Sheep, data=Datos, binomial) NSG <- glm2(Preys ~ Ndog+Location+Sheep, data=Datos, binomial) SPG <- glm2(Preys ~ Location+Food+Sheep, data=Datos, binomial) APG <- glm2(Preys ~ Hens+Food+Sheep, data=Datos, binomial) ASG <- glm2(Preys ~ Hens+Location+Sheep, data=Datos, binomial) ANG <- glm2(Preys ~ Ndog+Hens+Sheep, data=Datos, binomial) NAS<- glm2(Preys ~ Ndog+Hens+Location, data=Datos, binomial) ASP <- glm2(Preys ~ Hens+Location+Food, data=Datos, binomial) NAP <- glm2(Preys ~ Ndog+Hens+Food, data=Datos, binomial) NSP <- glm2(Preys ~ Ndog+Location+Food, data=Datos, binomial) PG <- glm2(Preys ~ Food+Sheep, data=Datos, binomial) NG <- glm2(Preys ~ Ndog+Sheep, data=Datos, binomial) NA1 <- glm2(Preys ~ Ndog+Hens, data=Datos, binomial) SP <- glm2(Preys ~ Location+Food, data=Datos, binomial) AS <- glm2(Preys ~ Hens+Location, data=Datos, binomial) SG <- glm2(Preys ~ Location+Sheep, data=Datos, binomial) AG <- glm2(Preys ~ Hens+Sheep, data=Datos, binomial) AP <- glm2(Preys ~ Hens+Food, data=Datos, binomial)

NP <- glm2(Preys ~ Ndog+Food, data=Datos, binomial) NS <- glm2(Preys ~ Ndog+Location, data=Datos, binomial) N <- glm2(Preys ~ Ndog, data=Datos, binomial) A <- glm2(Preys ~ Hens, data=Datos, binomial) P <- glm2(Preys ~ Food, data=Datos, binomial) S <- glm2(Preys ~ Location, data=Datos, binomial) G <- glm2(Preys ~ Sheep, data=Datos, binomial) Int<- glm2(Preys ~ 1, data=Datos, binomial)

PreysMod<- list(NASPG,NASG,NAPG,NSPG,ASPG,NPG,NSG,SPG,APG,ASG,ANG,NAS,ASP,NA P,NSP,PG,NG,NA1,SP,AS,SG,AG,AP,NP,NS,N,A,P,S,G,Int) Name <- c('NASPG','NASG','NAPG','NSPG','ASPG','NPG','NSG','SPG','APG','ASG','ANG','NAS', 'ASP','NAP','NSP','PG','NG','NA1','SP','AS','SG','AG','AP','NP','NS','N','A','P','S','G','Int') aictab(PreysMod, modnames = Name) global<-dredge(NASPG) a<- model.avg(global) summary(a) a$coef.shrinkage exp(a$coef.shrinkage) confint(a) exp(confint(a))

CHAPTER 3

FINE-SCALE MOVEMENTS OF RURAL DOMESTIC DOGS IN SENSITIVE

CONSERVATION AREAS IN THE SOUTH AMERICAN TEMPERATE FOREST

INTRODUCTION

Despite the global trend of urbanization, in most developing countries protected lands are still in remote areas far from industrialized and dense human settlements (Joppa and Pfaff

2009). The landscape surrounding these protected lands typically is inhabited by rural human communities that are highly dependent on local natural resources (Baland and

Platteau 1996, Shackleton et al. 2002). In these communities, proximity to protected lands can lead to conflict with conservation efforts (e.g. hunting bushmeat) but also can provide opportunity for the community (e.g. ecotourism) (Naughton-Treves et al. 2005,

Wittemyer et al. 2008, Andam et al. 2010). Agriculture and forest resources sustain the economy of these communities and they typically raise domesticated farm animals such as cattle and sheep as a protein source (Waters-Bayer and Bayer 1992, Dovie et al. 2006).

In forested areas, grazing areas are often created directly by farmers (Lambin et al. 2001) or indirectly by the grazing of their animals which prevents forest regeneration (Veblen et al. 2003, Steinfeld et al. 2006). In these rural communities, domestic dogs are common and play a variety of roles for the farmers, including the guarding of livestock and the

49 protection of households (Rigg 2001, Gehring et al. 2010a, González et al. 2012)(Chapter

2).

However, the presence of domestic dogs in these rural areas can be problematic.

Domestic dogs represent an increasing problem for biodiversity conservation (Young et al. 2011, Hughes and Macdonald 2013). Free-roaming dogs are a particular problem in developing countries (Dalla Villa et al. 2010) and are frequently found around or inside protected areas in Africa, South America and Asia (Butler and du Toit 2002, Butler et al.

2004, Fiorello et al. 2006, Koster 2008, Atickem et al. 2009, Lacerda et al. 2009, Vanak and Gompper 2010). The presence of domestic carnivores in natural ecosystems can affect the vertebrate local fauna through a variety of mechanisms (Vanak and Gompper

2010, Hughes and Macdonald 2013). Dogs, particularly free-roaming dogs, have been responsible for declines in threatened populations due to predation in a number of species including marine iguanas ( Amblyrhynchus cristatus ) (Kruuk and Snell 1981), Galapagos giant tortoises ( Chelonoidis nigra ) (MacFarland et al. 1974) and kiwis ( Apteryx mantelli )

(Taborsky 1988). Furthermore, the transmission of infectious diseases from high density populations of domestic dogs around parks also poses a significant threat to endangered wildlife (Cleaveland et al. 2000, Acosta-Jamett et al. 2011). For example, canine distemper and rabies from domestic dogs have caused population declines in African wild dogs ( Lycaon pictus ), lions ( Panthera leo ) and Ethiopian wolves ( Canis simiensis )

(Alexander and Appel 1994, Munson 1994, Roelke-Parker et al. 1996, Sillero Zubiri et al.

1996, Funk et al. 2001).

A critical gap in minimizing the risk to wildlife associated with interactions with domestic dogs is an understanding of the movement patterns and habitat use of these free- roaming dogs, particularly in relation to sensitive conservation areas and threatened species habitat. Previous telemetry studies in domestic dogs show clear selection of anthropogenic dominated landscapes (Vanak and Gompper 2010, Woodroffe and

Donnelly 2011). In particular, it is clear from this work that dogs mostly stay close to their house of origin. However, the existing studies involve infrequent radiotracking locations and, therefore, little information is available detailing how dogs move on the landscape during rarer foray behavior when they move away from houses. In order to better understand how dogs move during infrequent foray events, when they are more likely to interact with wildlife, movements need to be intensively monitored to capture all movement during a minute-hours timescale. To achieve a better understanding of foray behaviors in dogs, we used a new, frequent location radio tracking technology to study dog movements into sensitive conservation habitats in the Valdivian Temperate Forest in southern Chile. Our specific goals were to determine: 1) dog movements in relation to their house of origin, 2) habitats dogs utilize when exhibiting foray behavior, 3) the daily pattern of dogs forays and 4) the frequency of dog movements in and around protected areas or threatened species habitats in relation to particular land use patterns. Ultimately, the goal of this work is to inform and improve management of dog-wildlife interactions around protected areas while also decreasing disease transmission risk between dogs and endangered wildlife in the threatened Valdivian Ecoregion.

METHODS

Study area

The study area is located in the Coastal Range of Southern Chile in the Valdivian

Temperate Forest (39 ° 58’ Lat., 73 ° 31’ Long.). Rural communities living in the lowlands of the forest ecosystem utilize cleared lands for livestock production. Livestock in the study area included cattle and sheep located in pastures and chickens kept near households. Free-roaming domestic dogs were present in most households and were used primarily for farm animal protection (Chapter 2). Dog movement patterns were studied in Cadillal Alto, a small rural community that includes11 families/households located on the north bank of the Chaihuín River between two protected lands, the Valdivian Coastal

Reserve (VCR) and the Alerce Costero National Park (ACNP) (Fig. 1). This community is primarily accessible by boat, has no electricity and is geographically isolated from other rural communities. Grazing pastures for livestock are distributed predominantly in the northern lowlands of the Chaihuín River (Fig. 1). All families had dogs for a total of

21 domestic dogs in the community.

Dog sampling and data collection

Prior to the start of the study in December 2009, the goals and approaches for this research were presented to the community of Cadillal Alto and written permission obtained for all participating households. All available dogs in the village were enrolled in the study save those that were too aggressive to collar.

Dog movement information was collected using GPS radio collars developed for tracking hunting dogs (GPS/receiver Garmin Astro 220, Garmin, USA). Once placed and activated, the collars are capable of sending locations for up to 10 dogs at a time every 30 seconds-10 min for 12-24 hours depending on the number of dogs tracked and collar battery life. Locations were sent to a base station receiver that has a maximum reception range of 11 km. After the collars were fitted on a maximum of ten dogs in the village, the receiver was placed in a centralized location with a maximum distance of 1.5 km from all monitored households. Accuracy of GPS locations was measured for all radiocollars at the beginning and the end of the study with an error of < 20 meters from a stationary location. These cohorts of collared dogs were monitored continuously for 12-

24 hours, considered a monitoring period. The number of dogs monitored per period varied depending on how many owners and dogs were at home when the collars were being placed. Monitoring periods were repeated with an average 3.1 monitoring periods per month (range: 1-7) from January-June and November- December 2010 based on

53 community accessibility. No data were collected between July-September 2010 due to the seasonal heavy rains that limit access to the study site.

Habitat classification and analysis of dog habitat use and foray behavior

To determine habitat types, a land cover map was created by photo-interpreting in

ArcGIS 10.0 (ESRI Inc., Redlands, CA, USA) a satellite image from Google Earth of

May 2010 (QuickBird, Special Agency Digital Globe, USA), including 60-cm/pixel resolution and 4 multispectral bands. Habitats were categorized (native forest, eucalyptus, pastures, wetlands, rivers and trail/roads) at a minimum resolution of 3 x 3 meters.

To determine the overall pattern of dog movement, the percentage of locations per dog for each sequential 200 m interval from the household of origin was calculated. To analyze foray behavior in dogs, forays were defined as directional movement of a dog away from the house of origin for a minimum distance of 200 m from the household and for a minimum duration of 1 hour (Fig. 2). For all foray analyses, only locations within a foray (foray locations) were analyzed; all other locations were eliminated from analysis.

Data locations for monitoring periods were not independent considering the high level of detail (30 seconds-10 minutes), but for our purposes we required this level of detail in order to identify forays.

To determine dog habitat use during forays, all foray locations were subjected to a home range analysis, hereto referred to as ‘foray range’. The foray range was estimated using a

95% fixed kernel. Kernel estimators are based on probability “kernels”, which are regions around each point location containing some likelihood of animal presence

(Worton 1989). To estimate ‘foray ranges’ we used the Home Range Tools in the

ArcGIS 10.0. Foray range was calculated for all dogs having >35 foray locations, based on thirty locations being considered an adequate minimum threshold for fixed kernel home ranges (Seaman and Powell 1996). Habitat use of the forays was estimated at second and third order of selection as defined by Johnson (1980). Specifically, second order selection related to the composition of the foray range habitats of dogs compared to the study area, where the study area was defined by the minimum convex polygon of all dog radiolocations pooled. Third order selection, on the other hand, compared the composition of actual habitat used (as defined by the proportion of dog foray locations in each particular habitat category) to the available proportion of habitats within the foray range. Habitat use for both levels was analyzed using compositional analyses (Aebischer et al. 1993) using the package adehabitatHS and function compana (Calenge 2006) in software R (R Development Core Team 2011). In short, the proportions of habitats used to those available were compared between the study area and the foray range and then between the foray range and the total dog foray locations using a Wilk’s lambda ( λ) distribution. If habitat use in this analysis was found not to be random, pairwise comparisons were performed. Zero values of proportions can be found in the available habitats when the third-order habitat selection is under focus. In this case, it may occur

55 that some habitat types are available to some animals and not to the others. Because third order analysis involves analysis of actual locations and, therefore, can have individual dogs that do not use a particular habitat (zero denominator), a weighted mean lambda was used in this analysis as described by Aebischer et al. (1993: Appendix 2) and estimate of

P-values performed using a randomization test. River habitat was eliminated from this analysis since dog’s use of rivers was clearly transitory, with few locations, and no two locations in succession.

To compare day versus night foray behavior, the total radio-tracking hours per dog were calculated for the day (8:00 – 20:00 h) and the night (20:00-8:00 h) and compared to the hours of day and night foray activity per dog. Daylight activity of forays was analyzed similarly to the habitat use analysis described previously, by comparing the proportion of hours during day and night used during forays vs. the proportion of hours during monitoring. All statistical analyses were used an α level= 0.1 considering the limited sample size.

Finally, foray locations were characterized as being within a protected area or as using a known threatened species habitat. Threatened or endangered species of concern known to be in the study area included the Southern river otter ( Lontra provocax ) (Endangered),

Guigna ( Leopardus guigna ) (Vulnerable), Darwin’s fox ( Lycalopex fulvipes ) (Critically

Endangered) and Southern pudu ( Pudu puda ) (Vulnerable) (IUCN 2013). For these species, habitats of concern are native temperate forest (Guigna, Darwin’s fox, Southern

56 pudu) and riparian habitat (Southern river otter). Species presence in these habitats and in this study area had been confirmed by previous studies using camera traps (Sepúlveda

MA, unpublished data). In particular, the presence of the critically endangered Darwin’s fox has recently been confirmed by ongoing studies in the area (Farias et al. personal observations).

Forays in sensitive conservation areas were classified according to the following four criteria: 1) Riparian Foray : defined as a foray with dog locations within riparian forest along the Chaihuín River (<20 meters from river shore). This habitat was used as a proxy to represent the Southern river otter primary habitat (Sepúlveda et al. 2007); 2) Forest

Foray : defined as a foray with dog locations inside of a continuous native forest patch

>50 meters from the border. This is the primary habitat for pudu, guigna and Darwin’s fox (Dunstone et al. 2002, Jiménez 2007, Silva-Rodríguez and Sieving 2012); 3) Reserve

Foray : defined as forays with at least one dog location within the limits of the VCR; 4)

Park Foray : defined as a foray with at least one dog location within the limits of the

ACNP. All remaining forays were classified as confined to human dominated landscapes.

RESULTS

A total of 19,224 locations were collected from 14 rural dogs in 8 households. Two households did not agree to participate (2 dogs were present in each of these houses) and a third house had one aggressive dog that was excluded from the study. After the fourth month of the study, the frequency of GPS data collection in the radio collars changed from reading every 30 seconds to reading every 10 minutes; the reason for this change is unknown. Because all dogs were affected similarly, and because the dog cohort monitoring periods and forays were scaled in hours rather than minutes, the 10 minute interval was not considered problematic, and analysis was not altered following the change. After the change in recording time interval, the collar locations were verified using hand held GPS units to assure accuracy. The total number of monitoring days for all dogs pooled was 70.1 days (mean days per dog: 5.01 ± 0.08 SE). For all dogs, the mean duration of continuous monitoring period was 13.4 ± 0.7 SE hours. Information about dog’s sex, age, number of periods by collared dog is in Supplementary Material,

Table 1.

Dogs were observed in a foray only 5.3% of the total monitored time; the rest of the observations were at locations < 200 meters from the household or represented longer distance movements of less than an hour in duration. Forays ( n=51) were observed in all except three dogs (Table 1). One of these dogs was only collared for one monitoring period of 18 hours.

Foray habitat use was not random at a second order of selection (foray range versus study area) ( λ= 0.13, df = 3; P-value< 0.005) with dogs selecting pastures for movement during forays more than forest or roads compared to what would be expected based on the study area habitat distribution. In contrast, dogs selected forest less often than expected based on study area forest distribution (Fig. 4 and Table 2). In the analysis of third order foray habitat use (all foray locations versus foray range), a different pattern was observed.

Overall, third order selection was significant ( P<0.1) with roads being the most selected habitat for actual use by dogs and both road and pasture showing more dog locations than would be expected based on foray range alone compared to forest and wetland

(Randomization test: λ= 0.15; P= 0.086) (Fig. 4, Fig. 5 and Table 2).

The monitoring periods per dog were almost equally distributed between day and night with 49.8 ± 2.6 % occurring during daylight (8:00-20:00). However, the frequency of dog forays was mostly diurnal, with 91 ± 4.8%of foray hours occurring during daylight

(λ=0.30; df = 1; P= 0.0005).

There were a total of 13 forays made by 6 rural dogs into sensitive conservation areas

(42.8% of total radio-collared dogs and 25% of all forays) throughout the study- monitoring period. Riparian forays and Reserve forays occurred 6 and 5 times, respectively. Forest forays and Park forays only occurred once each, and the use of trails by dogs was particularly important during these forays (Fig. 6). 59

DISCUSSION

This study characterizes the fine-scale movements of dogs as they move away from their house of origin. These foray behaviors have distinct characteristics. First, they are rare.

As previously reported, dogs spend most of their time within 200 m of their homestead.

And secondly, when forays do occur, movement is directional and follows primarily human-dominated landscapes with occasional excursions into protected habitats. Better understanding of foray behavior will improve understanding of dog-wildlife interactions and disease transmission risk around conservation areas.

As expected, when all locations were analyzed together, dogs spent most of their time near their house of origin. This supports previous studies by Woodroffe and Donnelly

(2011) that found that the radiolocations of rural dogs in Kenya are negatively correlated with distance from bomas (livestock corrals where people, livestock and dogs reside) implying that dogs mostly staying near their households. The maximum detected distance of the dogs in this study (4 km) was also similar to previous studies. For example, Woodroffe and Donnelly (2011) in Kenya reported 3.2 km and Butler et al.

(2004) in Zimbabwe reported maximum dog ranges of 6 km with people and 3 km when dogs were alone. The same association of dog habitat use and human settlements has been reported in agricultural lands in India (Vanak and Gompper 2010) and also in more urbanized areas in Spain (Pita et al. 2009, Ordeñana et al. 2010) and North America

(Maestas et al. 2003) confirming the general pattern of a strong dependence and

60 association of dogs to human houses. This confirms that vicinity to human houses is probably the most important factor of rural dog distribution and is most likely driven by the food and shelter provision obtained at these sites. If we consider only these global analyses of dog movement in relation to houses, we could conclude that expected negative interactions between rural dogs and wildlife should be restricted to areas immediately adjacent to houses where dog activity is concentrated.

However, further analysis reveals the importance of forays of these dogs away from their house of origin. If dogs are getting important benefits from their association with humans, it might be expected that their foray activity would also be associated with humans. Such an association with humans is supported by the diurnal pattern observed in dog forays in this study. This kind of movement during daylight was found by

Woodroffe and Donnelly (2011) in Kenya and was attributed to dogs moving with their owners during routine farm or herder activities. Rural dog nocturnal activity without human accompaniment has been described in Australia by Meek (1999), who radiotracked dogs in aboriginal communities. However, these dogs were not monitored when they were with humans, so the relative importance of these nocturnal events in terms of overall movement is not clear. Feral dogs, on the other hand, defined by Vanak and Gompper (2009) as those living completely independent of food provision by humans, have been shown to be very active at night, possibly to avoid human activity

(Scott and Causey 1973, Daniels and Bekoff 1989, Boitani et al. 1995). The current study did not discriminate between dog movement accompanying humans versus solo

61 movement. Considering that most dogs in this study were designated as ‘working dogs’, we expect that humans accompanied most of these dog forays, particularly in human dominated landscapes such as roads and pastures, so in those cases dog movements are not independent of human movements. Future studies should determine if dog foray behavior changes with human accompaniment versus solo forays. This question of how human accompaniment changes timing and location of foray behavior will be important information to manage dogs near conservation areas and minimize dog-wildlife interactions. For example, predation on wildlife would be more likely if the dog was moving solo, whereas, dogs chasing farm animal predators is generally promoted by dog owners as described in Chapter 2.

The diverse habitat of the study area, composed of a mixture of agricultural and forest habitats, provided an opportunity to analyze actual dog foray movement versus that expected based on existing habitat. This analysis demonstrated that dogs do not move randomly on the landscape. Even considering the limited monitoring time per dog in this study (5 days per dog) and the limited frequency of foray events (5.3% of total monitoring) significant habitat selection was seen in these dogs during forays. At a second order of selection, dogs on forays selected pasture over forest or roads over what might be expected from the study area habitat distribution. Probably this selection of pastures over forest or roads is due to the movement of dogs with people or their farm animals (cattle and sheep). A different pattern emerged with the third order of selection, where dog radiolocations were found more on roads and pastures and less on forests and

62 wetlands than would be expected based on the composite foray ranges. The selection of roads for movement by dogs is a particularly interesting finding in this study for two reasons. First, roads and trails were important not only to connect households in the community, but also, dogs were using trails and roads to access conservation areas and intact tracts of forest habitat (Fig. 5). Second, roads were preferentially selected over forests and wetlands even though roads are a small proportion of habitat in the study area

(Fig. 4).

This particular result, showing the preferential use of roads and trails, was only observable in the current study due to the collection of frequent radiolocations over a number of hours. This, in turn, captures not only general movement on the landscape, but also detailed information about short-term directed movements and how they differ from time near the house (Fig. 2 and Fig. 5). Traditional radiotracking technology, which would normally track a few positions a day, is not capable of identifying and analyzing short-term trajectories like the dog forays in this study.

In Kenyan arid lands Woodroffe and Donnelly (2011) assumed an isotropic (uniform in all directions) movement of dogs from bomas. The assumption of isotropic movement may be reasonable in an open Savannah habitat, but our study shows that this assumption is not applicable in all habitats. Dog movement in the current study was strongly selective based on habitat. Forest was a clear barrier to movement, showing significant negative selection at both the second and third order of selection. Wetlands, on the other

63 hand, restricted fine scale dog movements (third order) while dogs appeared to select roads at this scale. Clearly, therefore, landscape attributes shape the distribution of dog movements in this ecosystem. In an agricultural landscape in India, Vanak and Gompper

(2010) observed a similar positive association of dogs to bare ground/roads and human settlements, supporting the selective nature of dog movement in heterogeneous landscapes. The frequency of forays into sensitive conservation areas appeared to depend primarily on accessibility to the park and reserve based on this selective use of habitat.

Forest represented an important barrier to dog movement into these areas, probably due to the dense understory vegetation dominant in this type of ecosystem (Alaback 1991).

These results not only demonstrate how particular habitats can limit dog movements but also highlight how the landscape composition around protected lands can influence the access of dogs to these areas. The habitats surrounding the VCR were composed primarily of pastures and the Chaihuín River, the latter of which clearly limited dog visits to the Reserve (Fig. 5). However, given the high frequency of dog movements in surrounding pastures adjacent to the river, it is not surprising that dogs were crossing the river and moving into the VCR. Forays that were made into habitats of threatened species had a similar selective use of habitats, with few forays in dense and continuous forest patches. In contrast, riparian forest habitats were more frequently visited, likely due to the pastures adjacent to these habitats. Because of the low number of forays observed and the small amount of total monitoring time, additional studies should be conducted to further refine these movement preferences. We observed 43% of collared dogs made at least one foray into a sensitive conservation area during an average

64 monitoring period per dog of 5 days ( n dogs =14; from a total dog population of 21 dogs). Although only 13 forays had dogs entering conservation areas, given the number of dogs in and around sensitive areas in Chile, the impact could ultimately be large.

Understanding the effects of dogs on wildlife is an urgent need for conservation (Young et al. 2011). However, the factors that influence these dog behaviors are rarely studied, and without this knowledge, dog management strategies that are compatible with the needs of rural communities are likely to be misguided. Rural dogs play important roles within rural settlements (Chapter 2) (Butler and Bingham 2000) including the guarding of households and livestock, and these functional roles need to be considered when discussing dog management. The promotion of management measures to restrict dog movements, particularly when owners do not supervise dogs, through means like leash use or kennels around farm animal enclosures would allow dogs to effectively perform their primary guarding duty, while also minimizing roaming and hunting behaviors.

Considering the type of forest ecosystem present in our study area and given that the movement of dogs into the protected areas was facilitated by the ease of movement through certain habitats, conservation programs could reduce dog access into forest ecosystems by reducing the presence of trails, particularly those close to rural communities. Such management would need to incorporate the opinions of the local human population to avoid conflict with the neighboring communities. Other measures such as riparian forest restoration, increasing forest patch sizes, increasing understory vegetation density and covering bare ground areas could decrease dog movements into

65 sensitive areas but could also have some benefits for farmers lands management (i.e. water quality, decreasing erosion by livestock or recreational use) (Loomis et al. 2000).

Finally, our study highlighted the importance of the understanding of small-scale movements, fine detail behaviors and rare events, which can improve our knowledge of how to prevent potential impacts of dogs around conservation lands and interactions with threatened species. This information was also adequate to provide management recommendations affordable at a conservation unit scale. Considering the low number of studies covering the study of small-scale movements future research on this topic is urgently needed.

Tables & Figures

Table 1. Summary of movements of 14 radio-collared rural dogs around the Valdivian Coastal Reserve, Alerce Costero National Park and Chaihuín River by number of forays detected, percentage of forays from total monitoring period and presence of forays in sensitive conservation sites and human dominated land uses.

Time spent Human dominated Dog Sensitive Conservation Forays Number in forays (% land forays ID Forays of total time) Riparian Forest Park Reserve D1 7.0 1 0 0 0 2 3 D2 5.7 0 0 0 0 1 1 D3 4.23 0 0 0 0 5 5 D4 0 0 0 0 0 0 0 D5 5.1 0 1 0 0 0 1 D6 11.1 1 0 0 0 9 10 D7 3.0 0 0 0 0 2 2 D8 0 0 0 0 0 0 0 D9 6.2 0 0 0 0 7 7 D10 4.5 0 0 0 0 2 2 D11 10.5 3 0 0 3 5 11 D12 6.5 1 0 0 2 3 6 D13 0 0 0 0 0 0 0 D14 6.0 0 0 1 0 2 3 Mean=5.3 Sum=6 Sum=1 Sum=1 Sum=5 Sum=38 Sum=51

Table 2. Habitat use selection of dog forays at second order of selection (foray range versus study area) and third order of selection (dog locations versus foray range). For second order analyses, a ‘+++’ indicates that the habitat in the row is selected more often for forays than the habitat in the column than would be expected given the distribution of habitat in the study area. For third order analysis, a ‘+++’ indicates that the habitat in the row was selected by dogs for forays more often compared to the habitat in the column than would be expected based on the foray range alone.

Forest Pastures Road Wetland Ranking Second order a

Forest 0 --- + - 1 Pastures +++ 0 +++ + 3 Road - --- 0 - 0 Wetland + - + 0 2

Third order b

Forest 0 ------0 Pastures +++ 0 - +++ 2 Road +++ + 0 +++ 3 Wetland + ------0 1 a λ=0.12; df = 3; P<0.005; b Randomization test: λ= 0.15; P= 0.086

Figure 1. Study area with locations of the 11 households that included dogs enrolled in this research. Note the difference of habitat composition between VCR and ACNAP in comparison to the location of households, open habitats and forested areas respectively.

Figure 2. Track of a rural domestic dog during a continuous monitoring period in relation to the owner’s household. The graph depicts three distinct forays (a, b and c), with a foray defined as continuous observations of at least 200 meters from the household for at least 1 hour (locations above the dotted line or outside the red buffer around the household). For foray analysis, only locations within a foray were included in the analysis.

Figure 3. Mean ± SE percentage of total locations by distances from dog owner’s households per dog. Bars indicate average for all dogs monitored ( n=14).

Figure 4. Proportion of habitat composition: 1) for total study area estimated by MCP of all locations during forays; 2) for foray ranges estimated by adaptive kernel 95% averaged by dog; 3) for all radiolocations averaged by dog. Error bars represent respective standard errors.

Figure 5. Total GPS locations of 14 rural domestic dogs monitored in southern Chile. Note predominance of dog locations in human dominated land use habitats.

Figure 6. Photos of rural owned dogs ranging in the study area. A) Dogs swimming across the Chaihuín River and B) reaching the river coast of the Valdivian Coastal Reserve.

SUPPLEMENTARY MATERIAL

Table 1. Sex, age, owner, number of periods radio monitored and average mean duration for each period (hours) by 14 rural dogs around protected areas in Southern Chile. Dog ID Sex Age Owner Periods Mean hours period (SE) D1 M 3 A 9 13.2 (2.3) D2 M 2 B 2 16.8 (10.5) D3 M 3 B 15 11.3 (2.2) D4 M 1 C 5 12.3 (2.7) D5 M 0.4 C 1 19.4 D6 M 1 D 15 12.1 (1.9) D7 F 12 E 10 9.4 (1.5) D8 F 10 E 11 11.2 (2.0) D9 M 4 F 18 12.9 (2.1) D10 M 4 F 5 18.3 (4.7) D11 M 2 G 22 11.6 (1.7) D12 M 2 G 16 11.1 (1.8) D13 F 2 H 1 18 D14 M 2 H 10 10.3 (2.6)

CHAPTER 4

INVASIVE AMERICAN MINK: LINKING PATHOGEN RISK BETWEEN

DOMESTIC AND ENDANGERED CARNIVORES

INTRODUCTION

Infectious diseases are a growing threat for biodiversity conservation (Daszak et al. 2000,

Lafferty and Gerber 2002). In recent years, several lines of evidence strongly suggest that a wide range of species, from plants (Anderson et al. 2004) to bats (Blehert et al.

2009), and from amphibians (Skerratt et al. 2007) to large carnivores (Murray et al. 1999) are suffering important population setbacks due to the emergence or reemergence of infectious diseases. In general, anthropogenic change is thought to be an important driver of infectious disease emergence, driven in part by changing human-domestic animal- wildlife interactions at the borders of human-dominated landscapes (Daszak et al. 2000,

Dobson and Foufopoulos 2001, Woolhouse et al. 2005). Overall, most research to date has focused on how human encroachment into natural habitats and habitat degradation alters human-wildlife contact (e.g. bushmeat hunting, urbanization) (McKinney 2002,

Wolfe et al. 2005, Jetz et al. 2007). However, anthropogenic change extends far beyond timber, roads and urbanization. One anthropogenic driver that bears further consideration and investigation is the impact that invasive non-native species have on disease risk in vulnerable wildlife.

Non-native invasive species are spreading throughout the world (Vitousek et al. 1996,

Simberloff 2005). Historically, these invasive species have been cause for concern since they often out-compete native species for resources (Roemer et al. 2002, Bonesi and

Palazon 2007). However, introduced species can potentially carry diseases of concern to vulnerable wildlife when they occupy and move in natural habitats (Crowl et al. 2008).

They may, therefore, pose a risk for disease transfer to the wildlife populations. Well- documented cases include the transmission of parapox virus from invasive Grey Squirrels

(Sciurus carolinensis ) to Red Squirrels ( S. vulgaris ) in the United Kingdom (Tompkins et al. 2003), or the global fungal pandemic, Batrachochytrium dendrobatidis, in amphibians associated with the global trade of African clawed frogs ( Xenopus laevis ) (Weldon et al.

2004). In fact, disease threats to native wildlife are now considered an important concern in managing invasive species (Clavero and García-Berthou 2005).

One particular group of exotic species that can play an important role in infectious disease dynamics are domestic animals (Wolfe et al. 2007). Domestic animals can reach high densities given the human provision of food, shelter and health care (Diamond and

Ordunio 1997). This, in turn, means that they can be more efficient hosts for maintaining disease transmission and persistence in their populations than wildlife, which are often lower density. Therefore, domestic animals may act as amplifiers of infectious diseases and serve as a source of pathogen spill-over for diseases that could not otherwise be maintained by rare native species (Grenfell and Dobson 1995, Woodroffe 1999).

Pathogen spill-over can occur when domestic animals are in proximity to wild species

(Lembo et al. 2008). This spill-over can occur from wildlife to domestic animals in cases where wildlife are in artificially high abundance as with transmission of bovine tuberculosis between food supplemented deer and domestic cattle in Michigan (Schmitt et al. 2002) or the emergence of Hendra virus associated with the dramatic increase in urban flying fox colonies in Australia (Plowright et al. 2011). However, in areas where wildlife populations are declining or abundances are low and domestic animals are numerous, the spill-over risk reverses and pathogens maintained in the domestic animal populations can spill over into otherwise naïve wildlife populations (Wolfe et al. 2007).

When this situation occurs in areas of importance for wildlife conservation, domestic animals can be a potential risk for threatened species. Examples include Toxoplasma gondii from domestic cats to sea otters (Miller et al. 2002), rabies from domestic dogs to

Ethiopian wolves (Sillero Zubiri et al. 1996) or Pasteurella from domestic sheep to bighorn sheep (Foreyt and Jessup 1982).

Few studies to date have investigated the ecology of interactions and disease transmission between domestic, invasive and native species. In some cases, disease transmission may be direct as the introduced animal invades the native animal’s habitat. However, the arrival of invader species could also alter the dynamics of interactions within the host community and therefore have consequences in terms of the risk of multihost pathogen disease transfer. This is a particular risk when two species that are segregated by differential habitat use and have few interactions become connected by an invasive

78 species whose new niche overlaps the two otherwise unconnected populations. This, in turn, would create new multihost pathogen dynamics with the invasive species potentially acting as a bridging host between the two populations. In this paper we investigate how the introduction of an invasive species alters carnivore interactions, and thus disease transmission risk, in a novel carnivore assemblage composed of a native, an invasive and a domestic species.

One disease system that is particularly interesting for understanding both the risk of disease spillover from domestic animals to wildlife and the role that an invasive species might play in multihost pathogen disease ecology is Canine Distemper Virus in carnivores. Wild carnivores naturally maintain low population densities, particularly endangered carnivores, making disease persistence difficult in these populations. Canine distemper is a multihost pathogen that is known to effect all carnivore families and, at least in some systems, appears to require multi-host systems for persistence resulting in cyclic outbreaks in wild populations (Craft et al. 2008, Almberg et al. 2009, Craft et al.

2009, Almberg et al. 2010). In addition, although considered a pathogen that requires direct transmission based on controlled experiments, the ecology of the disease in wild populations indicates that transmission can occur between geographically and behaviorally segregated populations (for example seals and terrestrial carnivores

(Mamaev et al. 1995) or dogs and lions (Cleaveland et al. 2000) that are highly unlikely to be interacting directly, indicating alternative hosts, fomites or indirect transmission may be playing a role in CDV persistence.

We studied the dynamics of interactions and, therefore, risk of CDV transmission, between three sympatric carnivore species in southern Chile: domestic dogs ( Canis domesticus), invasive American mink ( Neovison vison ) and endemic Southern river otters

(Lontra provocax ). Our study was conducted in southern Chile, an area that is characterized by a temperate climate where the survival of CDV in the environment should be long. The southern river otter is classified as an endangered species by the

IUCN (Sepúlveda M. et al. 2008) with main threats in Chile related to habitat degradation by riparian forest lost and human disturbance (Medina 1996, Medina-Vogel et al. 2003,

Sepúlveda M. et al. 2007, Sepúlveda M. et al. 2009). Domestic free-ranging rural dogs in Southern Chile are common and are used for farm animal protection and household guarding (Chapter 2). Rural dogs spend most of their time in the peri-human habitat

(Chapter 3) and dog owners in the study area often promote the interactions of dogs with wild carnivores when they are perceived as preying upon or attacking farm animals

(Chapter 2). This, in turn, has the potential to promote domestic dog-wildlife interactions and the potential transmission of CDV between species. River otters, on the other hand, are entirely riparian and rarely move more than 10 meters away from their river of origin (Sepúlveda et al. 2007). Thus, dogs and otters would normally not have high interaction risk. Since the 1960’s, however, the American mink was introduced in

Chile and Argentina for fur-farming and escapees from this industry quickly established feral populations that are today widely distributed throughout southern South America

(from 39° S to Tierra del Fuego) (Jaksic 1998, Jaksic et al. 2002). Mink scat have

80 frequently been seen at otter latrine sites during dietary studies (Rodríguez-Jorquera and

Sepúlveda 2011) and monitoring surveys (Sepúlveda, personal observation) of Southern river otters. Similar observations have been reported by Medina (1997) and Fasola et al.

(2009) in Chile and Argentina, respectively. Mink have also been seen near farms in the region (Chapter 2). Thus, introduced mink have the potential to inhabit both aquatic and terrestrial habitats (Dunstone 1993) to act as a bridge for CDV transmission between dogs and otters.

CDV is a morbillivirus that infects most mammalian carnivores and is characterized by its high lethality (Appel 1987, Deem et al. 2000). In recent years, CDV has been recognized as an important threat to wildlife since outbreaks have been implicated in declines of threatened or charismatic species such as the gray wolf ( Canis lupus )

(Almberg et al. 2012), black-footed ferret ( Mustela nigripes ) (Williams et al. 1988),

African wild dog ( Lycaon pictus ) (Alexander and Appel 1994) or African lion ( Panthera leo ) (Roelke-Parker et al. 1996). CDV makes a good model disease in this system since both dogs and mink are known hosts of the virus, and the endangered Southern river otter, as a mustelid, is likely to be highly susceptible to the disease (Appel and Gillespie

1972, Gorham 1999). Transmission of CDV is thought to be primarily through direct contact. CDV is thought to be highly unstable in the environment (Greene and Appel

2006), but under test conditions, the virus is able to survive from a few hours at 25°C to

14 days at 5°C (Shen and Gorham 1980). Also, virus direct viral transmission is generally by oronasal secretion but all bodily fluids can carry the virus (Haas and Barrett

1996). In fact, it can reach high levels in urine (Elia et al. 2006, Saito et al. 2006) and feces (Acton 2007). In addition, short-term indirect transmission by fomites has been postulated as a source of transmission, for example in lions and hyenas sharing carcasses or virus survival in an underground burrow of the Black footed ferret (Thorne and

Williams 1988) but to our knowledge the relative importance of fomites in CDV transmission has never been studied.

Here we evaluated the potential risk of CDV transmission within a carnivore assembly where native, free-ranging invasive and domestic species coexist. We hypothesized that mink, given their habitat use, would serve as an effective bridge species thus increasing the risk of transmission of pathogens such as CDV from dogs to otters. We expected to observe that household dogs would interact more frequently with minks than otters and that otters would interact more frequently with mink than with domestic dogs. Moreover, given the overlap in mink and otter niches and behaviors related to territoriality and scent marking, we investigated the usage of otter latrines by mink to determine the risk of indirect transmission for CDV. To test these hypotheses, this study had the following objectives: 1) to determine the frequency of contact among the three species near households (dog habitat) versus near rivers (otter habitat); 2) to understand how mink selectively use otter habitat and how it relates to indirect disease transmission risk; and;

3) to determine if rural dogs and mink in the study system had been exposed to CDV.

METHODS

Study area

The study was conducted in the Chaihuín and Colún rivers in Southern Chile (39º52'S,

73º25'W) (Fig. 1) in the Valdivian Temperate Forest. The study area included the Alerce

Costero National Park (24,694 ha; ACNP) and the Valdivian Coastal Reserve (50,530 ha;

VCR). Near the border of these protected areas and along the Chaihuín River are the villages of Chaihuín (more connected to towns) and Cadillal Alto (more isolated) (Fig.

1). Average annual rainfall is 2,500 mm per year and annual average temperature is 12°C

(range: 8°-17°C) (Luebert and Pliscoff 2005).

Measuring inter-species interactions

To determine risk of interactions between species we assumed that if two species use the same space within the published survival interval for CDV in the environment (< 14 days) then CDV transmission is possible. Interactions among the three species were assessed at two locations: 1) in riparian habitats where otter are associated with the river’s edge (Sepúlveda et al. 2007), and 2) in the peri-farm habitat around households with domestic dogs (Chapter 2). Because domestic dogs spend most of their time near households (Chapter 3), we interviewed dog-owning households about dog-wildlife interactions. Oral informed consent was provided by an adult in each interviewed

83 household in a standardized format. The purpose of the survey was explained in Spanish to each interviewee, and consent was obtained by having the interviewee state that they agreed to participate. The use of this survey and approach was approved by the

Institutional Review Board at the University of Minnesota (1304E31141). Fifty-two households were surveyed (see Supplementary Material) to determine: 1) the proportion of dog-owners that observed at least one mink or otter around their households during the previous year, and 2) whether the dog-owner observed at least one interaction between dogs and mink or dogs and otters. Dog-mustelids interactions reported by farmers were defined as those where the observer identified an actual contact such a dog killing or harassing mustelids.

To determine if mink or dogs were interacting with river otters in riparian habitat, we used camera traps to compare carnivore visits to otter latrines with visits to random sites at similar distances from the edge of the river. Interactions in this case were defined as two different species using the same camera trap site within a time interval between visits

< 14 days. We compared random with latrine sites to determine if dogs and particularly minks, considering our previous observations of minks using otter latrines, were selecting latrines over random sites. We consider otter latrines a high risk of indirect CDV transmission since otters concentrate scent marking at those sites with risky behaviors such as sniffing, rolling in the ground, defecating, urinating or marking with anal glands.

We deployed 91 camera traps between December 2010 and April 2011 (Bushnell trophy cam, Bushnell Corporation and Capture IR, Cuddeback) along the Chaihuín River ( n=

46) and Colún River ( n= 45) covering a total of 20 km of river for a total of 16 days per camera trap. Camera traps were set on all latrine sites (n=37) located through an intensive survey of the riverbanks of the two rivers (minimum distances between latrines: 55 m and

69 m for Chaihuín and Colún rivers respectively). Because otters spend most of their time within 10 m of rivers (Sepúlveda et al. 2007), 54 cameras were placed in randomly generated locations but within 5 m from the shore and with at least 200 m between camera traps, including distance to latrines. At each camera site we visually measured the percentage of understory (<1 m height) within a 5-m radius pivoting around the point with most otter signs (feces, anal secretions) (Depue and Ben David 2010), for random sites the radius center was the same where the camera was pointing out.

Mink site selection and disease transmission risk

To determine how mink selectively use otter habitat, we used a negative binomial regression where the number of mink visits to a site was the response variable. Because some mink visits were composed of several minutes of activity at the same trap site, we only considered independent events those that were three hours apart in the same site.

We used the time of camera monitoring (days) as an offset because the duration of monitoring period varied by camera. We used four predictor variables to determine the probability of a mink visiting a particular site: type of site (latrine or random), understory vegetation (%), number of otter visits and river site (Chaihuín River or Colún River). We used the number of otter visits to a site because we predicted that mink visits could be

85 negatively affected if the particular site is also used by an otter (i.e. intraguild killing)

(Donadio and Buskirk 2006, Simpson 2006). Understory vegetation and river site

(Chaihuín or Colún) were included as habitat covariates. For model selection purposes we used a backwards-stepwise selection process starting with a global additive model.

We assessed the fit of the best model using Chi-square goodness of fit.

To determine the number of interactions with potential risk of CDV transmission between species we quantified the number and time interval between visits of different species at each camera site. Mean, median and range of days between visits were calculated in the following indirect interactions between species: mink otter, otter mink, dog mink, dog otter and otter dog. The arrow symbol indicates a directional interaction at a single site where the animal to the left of the arrow visited the site first followed by the animal to the right of the arrow.

CDV serology

Mink and dogs were tested for CDV antibody titers to specifically characterize the CDV disease transmission risk to otters. Between November 2009 and February 2010

American minks were captured at the Chaihuín River (Fig. 1) using 15-30 cage traps

(Tomahawk ®, WI, USA) placed along a 10 km stretch of riverbank and baited with mackerel fish or chicken and checked twice per day for mink capture. All traps were

86 located within 10 m of the shore with at least ~ 200 m between traps. Animals were sedated in the trap with a combination of 10 mg/kg of Ketamine (Ketamil ®, 100 mg/ml,

Ilium, Troy Laboratories, NSW 2164, Australia) and 0.025 mg/kg of Medetomidine

(Domitor ®, 1mg/ml, Pfizer, Madrid 28002, España) i.m. Once sedated, a blood sample of

1 ml was drawn from the jugular vein and placed in a serum-separating tube and held at room temperature until processing. Animals were implanted with radiotracking devices at the time of sampling for a separate study. This ensured that animals were not resampled. After sampling and data collection the anaesthesia was reversed with the antagonist ATI (0,125 mg/kg i.m., Antisedan ®, Pfizer, Madrid 28002, España) administered i.m. We considered mink to be juveniles when they had adult teeth without abrasion and tartar, and adults otherwise (Fournier-Chambrillon et al. 2004).

Blood was drawn in March 2010 from at least one adult (>5 months) dog from each household in Chaihuín and Cadillal Alto villages. Any households reporting the use of canine vaccinations in the previous two years were excluded as well as any dogs <5 months old to avoid false positives due to vaccine-induced or maternal antibodies. Dog identification and age were provided by owners. Dogs were manually restrained, 2 ml of blood drawn from the cephalic vein and the blood processed similarly to the minks.

Owners were also asked to characterize dogs into one of two categories: unrestricted

(100% free roaming) or restricted otherwise.

Dog and mink samples were transported to a field laboratory and centrifuged within 6 hours of collection. Both mink and dog serum samples were stored in a -18°C freezer in the field laboratory and then transported on ice packs to the U.S. and stored at -70°C in a freezer prior to submission to the Veterinary Diagnostic Laboratory at Colorado State

University. Seropositivity to CDV was analyzed using a microneutralization test (Appel

1987), a test frequently used for domestic and wild carnivores (Almberg et al. 2009,

Gowtage-Sequeira et al. 2009, Prager et al. 2013). Because the CDV serum test was not validated for mink we compared two cut-off values of 1:8 and 1:16 based on values used in previous studies (Almberg et al. 2009, Gowtage-Sequeira et al. 2009, Prager et al.

2013). For the dog serum samples we used the recommendation of the laboratory, and a titer of ≥1:4 was considered positive.

The association of seropositivity of CDV with identified risk factors was determined for dogs for the following variables: (juvenile: defined as ≤ 2; otherwise adult), site

(Chaihuín or Cadillal) and dog management category (100% free-roaming or unrestricted otherwise unrestricted). Risk factors were determined using logistic regression (Agresti

2002). We assessed the goodness of fit of the model through the unweighted sum of squares test (Hosmer et al. 1997). Risk factors associated with CDV in mink were not determined because of the limited sample size ( n=23).

RESULTS

During the year prior to the interviews, more than 20% of respondents observed mink close to their households. Of these mink sightings, 42% involved at least one dog-mink interaction (Fig. 2). One dog-mink interaction resulted in the mink being killed by the dogs while the rest of the time the mink escaped from the dogs. This is in contrast to otters, where only 5% of owners reported seeing an otter and no dog-otter interactions were reported (Fig. 2).

Minks, otters and dogs were detected at 30, 17 and 3 camera sites, representing 97, 31 and 4 independent visits to these sites, respectively (Fig. 3). Fifty-three percent of the mink site visits that were captured on camera occurred at latrines. The likelihood of mink visiting camera sites was significantly ( P<0.05) influenced by site type (latrine/random) and river system (Chaihuín/Colún) based on a positive and significant incidence rate ratio

(Table 2). Mink were more than 4 times more likely to visit a latrine site/day of camera trapping and nearly 5 times more likely/day to be seen at the Chaihuín camera trapping sites versus the Colún sites (Table 2). Habitat site characteristic (% understory vegetation) and numbers of otter visits did not significantly influence mink visits to camera trap sites based on the negative binomial regression model.

All otter detections occurred at latrine sites (Fig. 3). Co-use of sites by carnivores was detected at 10 camera sites. Of these, otters and minks were co-using 8 camera sites, all

89 latrines, representing 22.2% of total latrines and 47% of visited latrines by otters (Fig. 5).

Mink and dogs co-used 2 sites: one random and one latrine (Fig. 4). In contrast, there was no co-use of any camera trap sites by dogs and otters, even though a dog was detected at a latrine site (Fig. 4). Cross-carnivore interactions were detected at camera trap sites with time intervals between otter mink, mink otter and dog mink (Table 3.

In addition, the median and average time between these interactions ranged from 1.2 to

1.7 days, well within the range of CDV environmental survival estimates (Table 3). No detections of mink dog or dog otter were observed.

Between 8 November 2009 and 16 February 2010 26 different minks were captured in

1,211 trap nights during 49 days effort. Of these mink serum samples, three were not analyzed due to low amount of serum obtained. Overall, CDV seroprevalence in mink was 39% (95% CI: 20-61) using cut-off 1:8 and 21% (95% CI: 8-44) using the 1:16 cut- off. Two mink cubs (< 3 months) presented positive CDV titers (1:16 and 1:512).

A total of 60 dogs were sampled for CDV testing. For dogs across all sites and ages,

CDV seroprevalence was 41.6% (95%CI: 29.3-55.0). Significant predictors of positive

CDV titers in dogs were age, with older dogs having higher ( P<0.05) CDV prevalence than younger dogs, and site, with dogs from Chaihuín having higher CDV seroprevalence than Cadillal Alto (Table 1). There was no evidence supporting an effect of roaming restriction (Table 1). Four seropositive juvenile dogs (<2 years old) had titers ranging from 1:16 - 1:1,024.

In addition to the information on interactions provided by interviews and camera traps, two mink carcasses were recovered by the research team that provide supplementary information (see Supplementary material, Fig. 1S and 2S). On September 2010, the carcass of a male adult mink was recovered 225 meters from a group of houses in

Cadillal Alto. The carcass was incomplete with a pattern of bruising, hemorrhage and associated crush injuries typical of large dog bites (see Simpson (2006)). Based on necropsy results and the proximity to households, the most likely cause of death was determined to be a domestic dog attack. The second mink carcass was found floating near an otter latrine on March 2011 at the Chaihuín River. The carcass was complete with clear bite wounds observed on the neck, nose and skull. At necropsy, the skull wounds corresponded to associated fractures of the parietal bone. Based on carcass location, and wound parameters, the most likely cause of death was determined to be river otter attack. No signs of consumption were detected.

DISCUSSION

This study shows, for the first time that invasive species can pose an increased disease risk to native species by acting as a ‘bridge host’ between threatened, rare native species and abundant, high transmission-risk domestic animals. In this system, dogs and otters were not seen to interact in either the peri-farm or the river habitat. In contrast, the

91 introduced American mink interacted with both carnivore populations in both habitats.

Of particular concern was the mink selection of otter marking sites. This observation merits further investigation as all studies of scent marking in carnivores indicate a within species communication process (Kleiman 1966, MacDonald 1980), but in other taxa, scent marking plays important roles in inter-species communication (Goulson et al. 1998,

Schulte 1998). In addition, camera traps were positioned to capture activity only in the center of latrines. Thus, there could potentially be more latrine interactions that were not captured in this study, making the presented results a conservative estimate of CDV risk.

Independent of the causes of the mink behavior, mink visits to otter latrines were common and within the temporal range of potential fomite based disease transmission.

This kind of indirect inter-species interaction could be an important contribution to disease ecology in multi-host pathogens in free-ranging wildlife.

We studied multi-host interactions in a forest ecosystem to determine risk of transmission of CDV to the river otter. From a conservation perspective, these results provide enough evidence of concern related to the potential impact of CDV in this threatened species.

Previous studies using mathematical modeling approaches have been able to quantify the relative risk of extinction under different scenarios for relatively well-studied species including Ethiopian wolves (Haydon et al. 2002), Saiga Antelope (Morgan et al. 2006) and Bighorn sheep (Cahn et al. 2011). The Serengeti lion CDV outbreak in 1994 and the persistence of CDV in Yellowstone National Park have also provided evidence of the importance of multi-host carnivore systems in maintaining CDV in wild populations

(Craft et al. 2008, Craft et al. 2009, Almberg et al. 2010, Craft et al. 2011). However, the most successful studies in understanding CDV risk to endangered species have historically relied on long-term field studies within probably the most well studied terrestrial ecosystems in the world. In these well-studied and relatively easily accessed systems, understanding of a complex disease like CDV is only just beginning to emerge.

This is in contrast to study areas like the dog-mink-otter system studies in this paper. In this system there is no/or limited information on wildlife (diseases or populations), the wildlife populations that exist are challenging to study in dense forest ecosystems, and the field sites are remote from road access with little infrastructure for research. Creating long-term research sites to better understand disease risk is both impractical and unlikely.

At the same time, the risk of negative effects of infectious disease to these rare and endangered populations are very real. It will be important to develop alternative approaches to monitor risk in these populations. Approaches such as those used in this study (farmer interviews, invasive species management, camera traps) can inform management policies at reasonable cost in both time and money.

Mink-other carnivore interactions were driven by different processes between the peri- farm and river systems. On the farm, interactions were driven by dog-mink conflict and dog-guarding behavior, as has previously been reported in Spain (Zuberogoitia et al.

2006). Along the river, mink showed clear preference for using otter latrines in comparison to random sites and there seemed to be a potential cross-carnivore marking behavior with mink marking closely following otter marking and vice versa.

Alternatively, both species could be selecting the same marking sites because of favorable habitat conditions for marking, likely the basis for the latrine site being established in the first place. This kind of cross-carnivore interaction merits further study as a mechanism of inter-carnivore disease transmission among endemic multiple – carnivore communities as well.

Mink in this study were more likely to be found in the Chaihuín river system. This is most likely related to differences in local mink abundances. The Chaihuín system is more associated with human habitation and is dominated by a mixture of agricultural lands and native forest compared to the Colún River, which is predominantly surrounded by native forest. As documented in this study, mink are capable of utilizing a mixed human-wildlife habitat, and this could potentially provide increased resources for the mink to exploit. Further research is needed to verify if this difference in trap visits does correlate to a difference in mink abundance and the causes of that difference.

This study shows that indirect contact is an important risk to the otter for disease transmission and that an introduced species can significantly alter this risk by acting as a

‘bridge host’ (Fig. 6). Mink are important in increasing that indirect transmission risk due to their niche in the system – moving between farm and river. We provided evidence that mink and dogs can interact by direct contact in aggressive encounters (i.e. mink killed or harassed by dogs around farms), but more interestingly, indirect interactions between otter and minks were frequently detected in otter latrines within a mean and

94 median of 2 days, an interval where pathogens such as CDV can persist and potentially be transmitted. This kind of co-use is comparable to other documented indirect interaction risks like sharing carcasses and could certainly lead to a risk of transmission through fomites in the environment. Mink are also much more abundant than native otters, probably related to their difference in body size (~1 kg the mink vs. ~10 kg the otter) (Dunstone 1993, Sepúlveda et al. 2007) and their adaptability to different habitats.

Their addition to this carnivore assemblage, therefore, also changes the local abundance of hosts for a multi-host pathogen like CDV and, therefore, increases the probability of pathogen persistence (De Castro and Bolker 2005). The particular similarities of otter and mink in relation to their use of habitats have been widely study in ecology (Ben-

David et al. 1996, Fasola et al. 2009, Medina Vogel et al. 2013, Valenzuela et al. 2013), but to our knowledge this study is the first to investigate how this niche overlap relates to multi-host disease transmission and persistence.

In relation to the CDV exposure observed in mink, to our knowledge, this is the first study documenting serological evidence of CDV in feral minks in South America. The level of CDV exposure with no observable clinical signs of distemper during captures is similar to the case observed in the domestic dog population sampled. Accepting the limitations of the use of serology in CDV we can only conclude that CDV is circulating in the ecosystem in both wild and domestic species. In terms of the time of exposure to the virus in mink, although an elevated titer (1: 512) was found in a mink cub (< 3 months), it is likely that the antibody detected is from a maternal origin (Chappuis 1998).

In the case of dogs, seroprevalence was associated with the more urban-accessible village, Chaihuín village, indicating that urban areas may be a source of distemper in this region as previously reported (Acosta-Jamett et al. 2011). Considering the high prevalence in adult dogs compared to juveniles and the absence of distemper clinical cases in dogs during sampling, it is likely that CDV exposure is historic rather than concurrent with the study. Although we did not sample otters, and as a consequence have no information regarding exposure to CDV in this endangered species, the high seroprevalence detected in mink, and the high contact rates between both species, suggest that otters are also exposed to this virus. Mustelid species of conservation concern have been reported to be highly sensitive to CDV declines (e.g. Black footed ferret) (Thorne and Williams 1988, Deem et al. 2000). However, studies describing CDV cases or outbreaks in mustelids are scarce probably due to the cryptic conditions of these species, their low abundances and the difficulties of detection in free-ranging species. In Florida, a CDV outbreak involved a threatened population of American minks ( Neovison vison evergladensis ) and gene sequences of virus shared close similarities with foxes in the same study area suggesting cross-species transmission (Cunningham et al. 2009). In the case of otters, CDV antibodies have been described in the American river otter ( Lontra canadensis ) (Kimber et al. 2000), and CDV-caused mortality in Eurasian otters and Asian clawless otter has been reported both in zoo and wild populations (Geisel 1979, Madsen et al. 1999, Mos et al. 2003, De Bosschere et al. 2005) indicating that presence of CDV is an actual threat to these species. In the ecosystem we studied only two previous accounts exist for CDV in wild carnivores: a clinical case of CDV in a chilla fox ( Lycalopex

96 griseus) (Gonzalez-Acuña et al. 2003) and the death of three radiocollared Darwin’s foxes ( Lycalopex fulvipes ) attributed to CDV (Jiménez et al. 2012). Considering the common diagnosis of CDV in urban dogs in Chile (López et al. 2009, Acosta-Jamett et al. 2011), including the nearby city of Valdivia (Ernst et al. 1987) to our knowledge this is the first study in the Valdivian Temperate Forest addressing the risk of CDV for this important ecosystem. Currently the mink have been present in South America for 50 years in this ecosystem and potential spread of CDV to otters could have occur and occurring but the absence of health and/or population monitoring of these species do not allow to determine population consequences on this species.

Conservation implications and management

Considering the conservation status of the Southern river otter we propose a precautionary management of CDV in this ecosystem based in our findings. Since domestic dogs appear to be an important and abundant host reservoir of CDV

(Cleaveland et al. 2000, Acosta-Jamett et al. 2011), it will be important for dogs to be vaccinated against CDV in areas of conservation concern (Cleaveland et al. 2006). This management will not only benefit otters, but also other species that presumably are being exposed to CDV including the highly endangered Darwin’s fox (Jimenez et al. 2012).

Approaches such as dog vaccination ring campaigns should be a priority particularly where otters and minks co-occur with large populations of dogs near major towns or cities and where CDV outbreaks are observed. Vaccination of otters or other endangered

97 carnivores is probably not feasible given the difficulties and dangers involved in trapping rare and secretive species. Similarly, mink vaccinations are probably not practical given the abundance of this species and high cost of trapping enough animals to induce sufficient levels of immunity in the population. Future studies should also provide recommendations on vaccine use since modified life virus vaccine in dogs has been implicated as a transmission risk to mustelids, that are highly sensitive to some viral strains (Greene and Vandevelde 2012).

In addition to managing the dogs, we recommend monitoring and possibly controlling mink. Even if contact rates between otters and dogs are low in the study area, mink are likely to facilitate the transmission of CDV between dogs and otters therefore the decrease of mink abundances could reduce the number of susceptibles at the carnivore community and particular the presence of this host bridge species decreasing the risk of distemper in otters. In addition, mink could be used as pathogen sentinels to obtain a better understanding of CDV in this environment considering their higher abundances, ecological similarities with otters and their relative ease of capture and sampling compared to native species (Sepúlveda et al. 2011). If campaigns to control mink (e.g. culling) are implemented, managers should be aware of potential consequences since reductions in carnivore densities can disrupt territorial structures resulting in an influx of new individuals and potentially increasing inter-species interactions and disease transmission (Hutchings and White 2000, White et al. 2008).

Overall, this study reveals how an introduced species can alter multi-host interactions and act as a bridge vector between common domestic species and rare native wildlife. Cross- species behavioral scent marking was a particularly interesting finding that should be further investigated as a means of indirect disease transmission among species. In the future, understanding these kinds of indirect interactions among species and, in particular, introduced species will be important to better manage disease transmission risk in wildlife and domestic animals. We urge for a major understanding of the role of behavior in carnivore interactions and their consequences for multi-pathogen dynamics.

Tables & Figures

Table 1. Multivariable logistic regression results of risk factors associated with CDV antibodies in serum samples from owned dogs (n=60) using seroneutralization test in Chaihuín River, Chile. Goodness of fit for the model was adequate ( P: 0.54).

Positive Negative OR (95% CI) P-value

Location

Chaihuín 22 23 1

Cadillal 3 12 0.22 (0.04-0.85) 0.0408

Age

Adult 21 21 1

Juvenile 4 14 0.25 (0.06-0.87) 0.04

Dog roaming

Restricted 3 5 1

Unrestricted 22 30 1.31 (0.26-7.57) 0.7426

100

Table 2. Multivariable negative binomial regression to estimate the Incidence Rate Ratio of visits of American mink to a camera trap site ( n=91) using log of number of sampling days as an offset.

IRR (95% CI) P-value

Intercept 0.063 (0.021-0.192) 0.000

Type

Random 1

Latrine 4.096 (1.336-12.557) 0.014

River

Colún 1

Chaihuín 4.726 (1.54-14.509) 0.007

101

Table 3. Time intervals of different directional interaction between otter, minks and dogs detected at camera traps.

Mink Otter Otter Mink Dog Mink Otter Dog

Number of events 8 6 2 0

Average (days) 4.0 3.9 1.2 ---

Median (days) 1.7 1.5 1.2 ---

Range (days) 0.6-16.5 0.6-4 0.2-2.3 ---

102

Figure 5. Study area showing the location of the Colún and Chaihuín River where camera traps were set up, and location of the human settlements Chaihuín and Cadillal Alto, where domestic dogs were sampled for CDV testing. For the Chaihuín River we also collected serum samples in American mink. Inset map shows the location of the study area in relation major human habitations including Valdivia city.

103

Figure 6. Frequency of dog owners (n= 52) reporting observing American mink or Southern river otter near their households at least once during the last year (total bar) and whether they detected a dog-mustelid interaction during that observation (black).

104

Figure 7. Frequency of occurrence of river otter, mink and domestic dogs at camera traps at riparian habitats (n= 91) by otter latrine and random sites.

105

Figure 8. Co-occurrence of dogs and American mink at an otter latrine site. A) Dog visiting an otter latrine (time 0), B) mink visiting the same latrine (time interval 2.3 days). Latrine was located in the VCR (southern riverbank) with nearby human settlements on the northern riverbank. Domestic dogs are able to cross the river and use the protected area as shown here.

106

Figure 9. Co-occurrence of American mink and Southern river otters at otter latrine sites. A) Mink visiting otter latrine (time 0), B) otter visiting same in latrine in A (time interval 2.7 days). C) Mink visiting otter latrine B (time 0), D) otter visiting latrine B (time interval 1.3 days). E) Mink visiting a otter latrine C (time 0), F) otter latrine C(time interval 1.4 days). Note picture C) shows a mink scent marking.

107

Figure 10. Schematic representation of the ‘bridge host’ concept in a multi-host pathogen system. A) Represents a host community where the target and reservoir host are segregated and therefore there is a low probability of transmission to the target hosts. B) Represents the same host community but adding the ‘bridge host’, the new host giving its common overlap with the previous segregated hosts is able to facilitate transmission from the reservoir to the target host.

108

SUPLEMENTARY MATERIAL

Dog’s owners questionnaire

1. ¿Cuantos perros tiene en este momento? (hembras, machos). (At present, how

many dogs do you have? males and females).

2. ¿Qué edad tienen? ( How old are they?).

3. ¿Cual es el origen de sus perros? (Where did you get them?).

4. ¿Como maneja a sus perros? permanentemente libre, amarrado, libre solo de día.

(How do you manage your dogs? Continuously free, leashed, leashed only during

the day ).

5. ¿Están vacunados o desparasitados? ( Are they vaccinated or treated against

parasites? ).

6. ¿Ha visto alguna vez alguno de estos animales? (Se mostraron 2 fotos para

identificacion de cada especie silvestre) ¿donde? ¿cuando? ¿que hacían? ¿qué

hizo usted? ( Have you ever seen one of these named species? [Two pictures for

each wild species were provided for identification] Where? When? What did they

do? What did you do? ).

109

7. ¿Ha cazado o atacado su perro alguna de las siguientes especies en el último año?

cuantos, donde, y cuando: Pudu, zorro, guiña, vison, chingue, huillin, coipo,

liebre. (During the last year, has your dog killed or attacked any of the following

species? how many, where and when: Pudu deer, foxes, lesser grisson, mink,

guigna, skunk, river otter, marine otter, coypu. ).

110

Supplementary figures

Figure 1S. Adult male mink carcass of a radiomonitored individual found 225 meters from a group of 6 households at Cadillal Alto and detected by radiotemetry with mortality sensor active. Left picture shows the general condition of the carcass. Right picture shows details of a bite wound. After necropsy most probable cause of death was attributed to domestic dogs.

111

Figure 2S . Mink female carcass detected at an otter latrine. Wounds observed were clean and only located at the head. Lower pictures show fracture of parietal bone as probable cause of death. Given location and kind of bites the killing was attributed to a river otter.

112

CHAPTER 5

GENERAL CONCLUSIONS

The three studies of this dissertation focused on understanding drivers of inter-species interaction in the context of disease transmission and conservation. Unlike most studies looking at disease transmission risk and species interactions, that tend to focus on quantifying and measuring interaction, these chapters seek to provide insight on why species are interacting; what impels otherwise solitary carnivores to come into contact with each other? In addition, the unique nature of the three-carnivore system studied allowed the research to investigate interactions not just among free-ranging populations, but among a domestic species (dog), invasive species (mink) and endangered free- ranging species (otter), providing insight on how species interact at the interface between humans and wildlife. Each study focuses on a different aspect of this interface: 1) how humans drive interactions between domestic dogs and wildlife; 2) how dog preferential use of the landscape alters interaction risk; and 3) how the behaviors of introduced mink alter interaction between otherwise isolated dogs and otters. These studies will provide essential information not only for managing interactions to minimize disease and conservation risk in this carnivore assemblage, but also for investigating and managing interactions in other multi-species systems.

113

Chapter 1 investigated how humans drive dog-wildlife interactions in rural Chile.

Specifically, it found that dogs are important to farmers as protectors of livestock in rural

Chile and, in that role, have a higher probability of dog-carnivore interactions.

Interactions with prey species, on the other hand were more likely to occur in dogs that were not provided with good quality food by their owner, indicating a hunting, food procurement behavior. In this study, an interesting finding that could assist management of the dog-wildlife interactions was the intensive management of dog demography by the owners. The population is highly skewed toward males and has a relatively low rate of reproduction with culling of female and pups being common. Possibly more interesting, the dog population seems to be maintained through the importation of male dogs from urban areas. This could provide an important mechanism for disease control and wildlife interaction management by providing farmers to access to vaccinated guarding breeds of dogs that would be less likely to exhibit hunting behaviors.

In Chapter 2, data show that rural dogs spend most of their time within 200 m of their own household. Even when they move away from their own household (forays), they preferentially use human-dominated and open landscapes (pastures) and trails/roads for movement and present a marked diurnal pattern of forays. Forests seem to represent barriers to movement, and, when entered, trails and/or roads were selected for movement.

Thus, protected areas with a buffer of forest had fewer visits by dogs than protected areas buffered by rivers or pastures, but considering the rarity of these events during our monitoring, further studies need to clarify those differences. Threatened forest species

114

(Darwin’s fox, Southern pudu and Guiña), therefore, are more prone to be exposed to dog interactions when near households or trails. Otters, on the other hand as inhabitants of riparian forest, are generally more exposed to dogs presence since riparian forests, where the otters live, are generally in lowland areas where pastures occur (Sepúlveda et al.

2007, Sepúlveda et al. 2009), habitats that are highly selected during dog forays.

Although this study did not examine dog-wildlife interactions per se, it does provide important information about how dogs interact with wildlife habitat, and, in particular, protected lands, providing insight into how dog behavior might drive wildlife interactions, and, in turn, how dog behavior such as movements and habitat can be potentially modified to decrease the risk of dog-wildlife interactions.

Finally, Chapter 3 investigated interactions and CDV exposure among all three carnivores in this system: dog, mink, and otter. This study found that the newly introduced and invasive American mink, through its particular habitat use, creates a significantly increased risk of disease transmission between domestic dogs and river otters. Whereas otters and dogs were not detected to interact, mink were frequently seen in both the dog and the otter habitats. In fact, the mink selected river otter latrines over random sites with such frequency, that mink and otter visits often co-occurred within the window of survival of CDV in the environment. Both dogs and mink were confirmed to be exposed to CDV in this ecosystem. Given the context of increased interactions generated by this new carnivore assemblage, it appears that CDV transmission is an important threat to the conservation of the endangered Southern river otter in this

115 ecosystem. It is clear from this work that the potential source of infection (dogs) and the target host species (otter) do not normally have significant habitat overlap. However the introduction of mink whose habitat use and behavior lead to frequent interactions with both of these otherwise segregated hosts, significantly increases the risk of disease transmission on this landscape. In this system, therefore, the mink can be defined as a

‘bridge host’ (Fig. 1).

Figure 1. Schematic figure represent the conceptual idea of the ‘bridge host’. A) Two segregated hosts with a low probability of pathogen transmission from reservoir to target host. B) Same situation as A) but adding the new host that overlaps habitats of the segregated species and therefore increase the risk of transmission from reservoir to target host.

116

Moreover, given the level of indirect interactions between mink and otters in latrines

(Fig. 2), it will be important be better understand the role that such indirect interactions play in the ecology of disease in free-ranging animals. Such inter-species interactions are not well characterized. Why are mink and otters apparently alternating visits at latrine sites? It will be important to not only understand the drivers of this kind of behavior, but also what are the implications for multi-host disease dynamics. In the case of CDV, although understood to be passed by direct transmission in domestic animal and laboratory systems, the number and diversity of outbreaks in wildlife seems to indicate that the situation is more complex in free-ranging and multi-host systems. What is the role of environmental transmission for this and other diseases where multiple species appear to be exhibiting territorial marking across species? Similarly, it will be important to understand the complexities of carnivore interactions in other similar ecosystems where the impact of distemper in these populations is well known. For example, what role does sharing carcasses between lions and hyenas or wolves and coyotes contribute to

CDV or other multi-host disease transmission? This question of how disease transmission is apportioned between direct and indirect transmission in multi-host disease systems has not been adequately addressed to date. Certainly the approach of only evaluating direct contact transmission is wholly inadequate to describe disease ecology in multi-host disease systems when indirect contacts can occur within the time window of environmental persistence (Fig. 2).

117

Figure 2. Schematic representation of transmission detections in a multi host framework: A) Detection of direct contact interactions between species 1 and species 2; B) Detection of direct interactions and indirect interactions between individuals same species in same situation. Tracking lines represent movement of duration within the viability time of persistence of the pathogen where indirect interactions actually represent risk of transmission. Note that the detection of direct contact between different species is inadequate to describe all the potential disease transmission interactions even in a system with only two species interacting.

Overall, these studies provide an interesting insight into the complexities of domestic animals, invasive species, wildlife interactions in terms of not only what interactions are occurring, but also why and how they are occurring in a context of conservation and infectious disease transmission. Based in the results obtained from our studies it appears that human behaviors, the effect of the landscape heterogeneity, the recent arrival of invasive species and animal behaviors such as scent marking and territoriality are all playing a role in driving inter-species interactions highlighting the complexities of the understanding of carnivore assemblages.

118

CONSERVATION MANAGEMENT IMPLICATIONS

In order to manage rural dog’s impact on vulnerable wildlife around protected areas it is important to have an adequate understanding of the role that rural dogs play for local communities. Ideally, knowing this information would allow management strategies that enhance rural community livelihood while also protecting wildlife. Based on the current studies, there could be such ‘win win’ scenarios in this community. For example, because most dogs in this community are imported from urban areas, it would be relatively simple to provide access to dog breeds that are selected for guarding rather than hunting behaviors. These dog breeds have been shown to minimize both dog-prey and dog-carnivore conflict in other rural areas (González et al. 2012). The importation also provides an opportunity to vaccinate them prior to introduction into this community, moreover, which would minimize disease risk not only to the dog population, but also to the sympatric carnivores in the assemblage. In addition, because hunting and foray behaviors are considered undesirable by dog owners as a distraction from their primary role as guards of livestock and household, measures to restrict dog movement through leashes or kennels might be acceptable to the community, thus reducing un-supervised dog movements. Given that most dog movement is near human habitation, this would also likely minimize dog attacks on livestock, a highly undesirable outcome of dog forays. In addition, if kennels were located near poultry houses, this would strengthen

119 the ability of dogs to protect livestock from predation of wild carnivores while also reducing the associated dog-carnivores interactions. As has been noted before by Silva-

Rodriguez and Sieving (2011), providing an adequate diet to dogs could also minimize dog forays and hunting behaviors thus minimizing interactions with wildlife, particularly prey species. One common measure to control dog populations in the past has been the use of surgical sterilization (Reece 2005). The studies presented in this dissertation make clear that sterilization would not be useful to control the study population here given the structure of the dog population (mostly males), although it could help eliminate the unnecessary killing of bitches and pups. Given the degree of isolation and the different priorities about domestic animal management in these rural communities, these interventions would be most successful with support from conservation organizations or government agencies. Considering the habitats used during dog forays, protected land or conservation managers could limit the access of dogs into sensitive conservation areas by limiting the number of road/trails within important forest patches or also by creating dense forest buffers around protected lands or riparian habitats as a deterrent to dog movement.

In addition to improving management of domestic dogs near conservation areas, this work also provided insight into the role invasive species play in changing inter-species dynamics on the landscape and management strategies to minimize this risk. As an invasive species, this work shows that the American mink could serve as a pathogen sentinels in this ecosystem due to: 1) its high abundance relative to native carnivores

120 leading to relatively easy access to individuals; 2) the acceptability and, even, necessity of trapping invasive animals in vulnerable areas as opposed to exposing endangered and rare native wildlife to the risks of trapping and anesthesia; and; 3) its bridging interactions among multiple carnivores which makes disease exposure in mink a marker of disease presence in the whole community. Also, as an invasive species, mink are likely to be the target of culling initiatives (e.g. predator control programs), which could present an important opportunity to monitor disease opportunistically. However, in planning culling of invasive species, precautions should be taken in managing disease since culling of individuals can have unintended consequences as evacuated territories can prompt new mink immigration, and, potentially, disease introduction with the new immigrants (Tuyttens et al. 2000).

This dissertation used a model assemblage of domestic, invasive and native carnivores to understand threats to endangered mammals in the Valdivian Temperate Forest associated with inter-species interaction and conflict. The Southern river otter was the major native species of concern in this study system; however other species in the same ecosystem such as Darwin’s fox ( Lycalopex fulvipes ), Guigna ( Leopardus guigna ) and Southern pudu ( Pudu puda ) will also benefit from the recommendations presented above. The lack of existing long-term monitoring of wildlife and the challenge of accessing information about wildlife in remote and often dense forest habitat is a particular problem for this ecosystem. Most long-term monitoring of wildlife has occurred in more open, accessible habitat where even species with low population density, like carnivores, can be tracked

121 visually on a regular basis (Almberg et al. 2012). In the current study, the combined use of traditional and novel tools including serology, camera traps, hunting dog collars and farmer questionnaires provided critical information about disease risk and interactions with the endangered otter populations without invasive and risky interventions in the otters. Such methods will be important in the future to better understand complex disease interactions in remote and challenging habitats.

Given the high prevalence of exposure to CDV in dogs and mink in this system, and considering high impact of CDV outbreaks on other carnivore populations in Africa or

North America (e.g. lions or black footed ferrets), it is urgent that Chile adopts a proactive carnivore monitoring strategy to protect its endemic and endangered carnivores.

In order to effectively manage these risks, the monitoring scheme would need to detect rapid decreases of populations or increases of infected individuals, especially considering the rapid impacts of CDV on vulnerable carnivore populations.

FUTURE RESEARCH RECOMMENDATIONS

The work accomplished in this dissertation provides important insight into potential future directions for research, not only in the carnivore assemblage studied in Southern

Chile, but for understanding drivers of inter-species interactions and how they relate to disease risk and wildlife conflict throughout other ecosystems. The studies open many

122 questions related to animal behavior and populations, humans-social aspects and management.

In relation to behavioral and populations studies, to the question of what is motivating the overlap of space and time between and within species is of particular relevance. In particular, how multi-host pathogens can persist in carnivore assemblages where interactions can be driven by avoidance (segregation) or attraction (overlap). How generalizable are the interactions or pathogen dynamics observed in Africa or North

America, where the presence of apex predators (i.e. lions or wolves) create carcasses that are sources of potential interactions with other mesopredators? How can CDV persist in other ecosystems where carnivore interactions are driven by other factors, as in the assemblage studies where apparent territorial interactions between species was a source of indirect exposure? How is CDV affected under different conditions of the environment or the carnivore community composition? How CDV persist in a carnivore assemblage without dogs as a spillover host, as is probably occurring in some areas of

North America?

The study of the causes for interspecies interaction processes, particularly marking behavior, is a topic that is poorly studied in the literature. Considering that similar situations to the association of mink to otter latrines could be occurring in other ecosystems, an obvious study system would be the introduced American mink in Europe.

Most likely the American mink in Europe, as in Chile, is sharing similar aquatic habitats

123 and possibly exhibiting similar co-marking behaviors with other threatened mustelids including the European mink ( Mustela lutreola ) and the Eurasian otter ( Lutra lutra ).

In the future, data generated here about interactions, population parameters and habitat use could be used to develop mathematical models to predict the risk of extinction of conservation target species, the control of invasive species, vaccination programs and other measures. This in turn could provide information about what parameters are most urgent to understand or what management scenarios are most effective to prevent CDV impacts in this multi-host system.

Finally, we emphasize the needs for long-term monitoring studies of multiple carnivore assemblages. To date, these studies have focused on open, easy to access populations.

This will require new techniques that allow monitoring not only the population trends, but also the rapid identification of rare events in hard to access habitats such as a CDV outbreak in forest wild carnivores.

124

REFERENCES

Acosta-Jamett, G., W. S. K. Chalmers, A. A. Cunningham, S. Cleaveland, I. G. Handel, and B. M. D. Bronsvoort. 2011. Urban domestic dog populations as a source of canine distemper virus for wild carnivores in the Coquimbo region of Chile. Veterinary Microbiology 152 :247-257.

Acosta-Jamett, G., S. Cleaveland, A. A. Cunningham, and B. M. D. Bronsvoort. 2010. Demography of domestic dogs in rural and urban areas of the Coquimbo region of Chile and implications for disease transmission. Preventive Veterinary Medicine 94 :272-281.

Acton, A. E. B. 2007. Evaluation of noninvasive molecular monitoring for fecal pathogens among free-ranging carnivores. University of North Carolina State.

Aebischer, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74 :1313-1325.

Agresti, A. 2002. Categorical data analysis. 2nd edition edition. John Wiley and Sons, Inc, Hoboken, New Jersey.

Alaback, P. B. 1991. Comparative ecology of temperate rainforests of the Americas along analogous climatic gradients. Revista Chilena De Historia Natural 64 :399-412.

Alexander, K. A. and M. J. Appel. 1994. African wild dogs (Lycaon pictus) endangered by a canine distemper epizootic among domestic dogs near the Masai Mara National Reserve, Kenya. Journal of Wildlife Diseases 30 :481-485.

Alexander, K. A., P. W. Kat, L. A. Munson, A. Kalake, and M. J. G. Appel. 1996. Canine distemper-related mortality among wild dogs (Lycaon pictus) in Chobe National Park, Botswana. Journal of Zoo and Wildlife Medicine 27 :426-427.

Almberg, E., L. Mech, D. Smith, J. Sheldon, and R. Crabtree. 2009. A serological survey of infectious disease in Yellowstone National Park's canid community. Plos One 4: e7042.

Almberg, E. S., P. C. Cross, A. P. Dobson, D. W. Smith, and P. J. Hudson. 2012. Parasite invasion following host reintroduction: a case study of Yellowstone's wolves.

125

Philosophical Transactions of the Royal Society B: Biological Sciences 367 :2840- 2851.

Almberg, E. S., P. C. Cross, and D. W. Smith. 2010. Persistence of canine distemper virus in the Greater Yellowstone Ecosystem's carnivore community. Ecological Applications 20 :2058-2074.

Andam, K. S., P. J. Ferraro, K. R. Sims, A. Healy, and M. B. Holland. 2010. Protected areas reduced poverty in Costa Rica and Thailand. Proceedings of the National Academy of Sciences 107 :9996-10001.

Andelt, W. F. 1992. Effectiveness of livestock guarding dogs for reducing predation on domestic sheep. . Wildlife Society Bulletin 20 :55-62.

Andelt, W. F. and S. N. Hopper. 2000. Livestock guard dogs reduce predation on domestic sheep in Colorado. Journal of Range Management 53 :259-267.

Anderson, P. K., A. A. Cunningham, N. G. Patel, F. J. Morales, P. R. Epstein, and P. Daszak. 2004. Emerging infectious diseases of plants: pathogen pollution, climate change and agrotechnology drivers. Trends in Ecology & Evolution 19 :535-544.

Appel, M., C. Reggiardo, B. Summers, S. Pearce-Kelling, C. Mare, T. Noon, R. Reed, J. Shively, and C. Örvell. 1991. Canine distemper virus infection and encephalitis in javelinas (collared peccaries). Archives of Virology 119 :147-152.

Appel, M., W. Shek, and B. Summers. 1982. Lymphocyte-mediated immune cytotoxicity in dogs infected with virulent canine distemper virus. Infection and Immunity 37 :592-600.

Appel, M. J. and J. H. Gillespie. 1972. Canine distemper virus. Springer.

Appel, M. J., R. A. Yates, G. L. Foley, J. J. Bernstein, S. Santinelli, L. H. Spelman, L. D. Miller, L. H. Arp, M. Anderson, and M. Barr. 1994. Canine distemper epizootic in lions, tigers, and leopards in North America. Journal of Veterinary Diagnostic Investigation 6:277-288.

Appel, M. J. G. 1987. Canine distemper virus. Pages 133-159 pp. in M. J. G. Appel, editor. Virus infections of Carnivores. Elsevier Science Publishers, Amsterdam, The Netherland.

126

Appel, M. J. G. and B. A. Summers. 1995. Pathogenicity of morbilliviruses for terrestrial carnivores. Veterinary Microbiology 44 :187-191.

Atickem, A., A. Bekele, and S. D. Williams. 2009. Competition between domestic dogs and Ethiopian wolf (Canis simensis) in the Bale Mountains National Park, Ethiopia. African Journal of Ecology 48 :401-407.

Atwood, T. C. 2006. Behavioral interactions between coyotes, Canis latrans, and wolves, Canis lupus, at ungulate carcasses in southwestern Montana. Western North American Naturalist 66 :390-394.

Baland, J.-M. and J.-P. Platteau. 1996. Halting degradation of natural resources: is there a role for rural communities? Food & Agriculture Org., Oxford.

Bangs, E. E. and S. H. Fritts. 1996. Reintroducing the gray wolf to central Idaho and Yellowstone National Park. Wildlife Society Bulletin 24 :402-413.

Barrett, T., I. Visser, L. Mamaev, L. Goatley, M.-F. Van Bressem, and A. Osterhaus. 1993. Dolphin and porpoise morbilliviruses are genetically distinct from phocine distemper virus. Virology 193 :1010-1012.

Bartlett, M. S. 1957. Measles periodicity and community size. Journal of the Royal Statistical Society. Series A (General) 120 :48-70.

Begon, M., J. L. Harper, and C. R. Townsend. 1996. Ecology: individuals, populations, and communities. Wiley-Blackwell.

Ben-David, M., R. T. Bowyer, and J. B. Faro. 1996. Niche separation by mink and river otters: Coexistence in a marine environment. Oikos 75 :41-48.

Bieringer, M., J. W. Han, S. Kendl, M. Khosravi, P. Plattet, and J. Schneider-Schaulies. 2013. Experimental Adaptation of Wild-Type Canine Distemper Virus (CDV) to the Human Entry Receptor CD150. Plos One 8:e57488.

Blancou, J. 2004. Dog distemper: imported into Europe from South America? Historia medicinae veterinariae 29 :35.

Blehert, D. S., A. C. Hicks, M. Behr, C. U. Meteyer, B. M. Berlowski-Zier, E. L. Buckles, J. T. Coleman, S. R. Darling, A. Gargas, and R. Niver. 2009. Bat white- nose syndrome: an emerging fungal pathogen? Science 323 :227-227.

127

Boitani, L., F. Francisci, P. Ciucci, and G. Andreoli. 1995. Population biology and ecology of feral dogs in central Italy. Page 217 in J. Serpell, editor. The domestic dog: Its evolution, behaviour, and interactions with people.

Bonesi, L. and S. Palazon. 2007. The American mink in Europe: status, impacts, and control. Biological Conservation 134 :470-483.

Brooks, T. M., R. A. Mittermeier, G. A. B. da Fonseca, J. Gerlach, M. Hoffmann, J. F. Lamoreux, C. G. Mittermeier, J. D. Pilgrim, and A. S. L. Rodrigues. 2006. Global biodiversity conservation priorities. Science 313 :58-61.

Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information–theoretic approach. 2nd edn edition. Springer-Verlag, New York, NY.

Butler, J. and J. Bingham. 2000. Demography and dog-human relationships of the dog population in Zimbabwean communal lands. Veterinary Record 147 :442-446.

Butler, J. R. A. and J. T. du Toit. 2002. Diet of free-ranging domestic dogs (Canis familiaris) in rural Zimbabwe: implications for wild scavengers on the periphery of wildlife reserves. Animal Conservation 5:29-37.

Butler, J. R. A., J. T. du Toit, and J. Bingham. 2004. Free-ranging domestic dogs (Canis familiaris) as predators and prey in rural Zimbabwe: threats of competition and disease to large wild carnivores. Biological Conservation 115 :369-378.

Cahn, M. L., M. M. Conner, O. J. Schmitz, T. R. Stephenson, J. D. Wehausen, and H. E. Johnson. 2011. Disease, population viability, and recovery of endangered Sierra Nevada bighorn sheep. The Journal of Wildlife Management 75 :1753-1766.

Calderon, M. G., P. Remorini, O. Periolo, M. Iglesias, N. Mattion, and J. La Torre. 2007. Detection by RT-PCR and genetic characterization of canine distemper virus from vaccinated and non-vaccinated dogs in Argentina. Veterinary Microbiology 125 :341-349.

Calenge, C. 2006. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling 197 :516-519.

Caro, T. and C. Stoner. 2003. The potential for interspecific competition among African carnivores. Biological Conservation 110 :67-75.

128

Carre, H. 1905. Sur la maladie des jeunes chiens. Les Comptes Rendus de l’Académie des Sciences 140 :689-690.

Chappuis, G. 1998. Neonatal immunity and immunisation in early age: lessons from veterinary medicine. Vaccine 16 :1468-1472.

Ciucci, P. and L. Boitani. 1998. Wolf and dog depredation on livestock in central Italy. Wildlife Society Bulletin 26 :504-514.

Clark, T. W. and T. M. Campbell III. 1981. Additional black-footed ferret (Mustela nigripes) reports from Wyoming. Western North American Naturalist 41 :360-361.

Clavero, M. and E. García-Berthou. 2005. Invasive species are a leading cause of animal extinctions. Trends in Ecology and Evolution 20 :110-110.

Cleaveland, S., M. G. J. Appel, W. S. K. Chalmers, C. Chillingworth, and M. Kaare, Dye, C. 2000. Serological and demographic evidence for domestic dogs as a source of canine distemper virus infection for Serengeti wildlife. Veterinary Microbiology 72 :217–227.

Cleaveland, S., M. Kaare, D. Knobel, and M. K. Laurenson. 2006. Canine vaccination— providing broader benefits for disease control. Veterinary Microbiology 117 :43- 50.

Cooper, S. 1991. Optimal hunting group size: the need for lions to defend their kills against loss to spotted hyaenas. African Journal of Ecology 29 :130-136.

Courtenay, O., R. J. Quinnell, and W. S. K. Chalmers. 2001. Contact rates between wild and domestic canids: No evidence of parvovirus or canine distemper virus in crab- eating foxes. Veterinary Microbiology 81 :9-19.

Craft, M., P. Hawthorne, C. Packer, and A. Dobson. 2008. Dynamics of a multihost pathogen in a carnivore community. Journal of Animal Ecology 77 :1257-1264.

Craft, M., E. Volz, C. Packer, and L. Meyers. 2009. Distinguishing epidemic waves from disease spillover in a wildlife population. Proceedings of the Royal Society B: Biological Sciences 276 :1777-1785.

Craft, M. E., E. Volz, C. Packer, and L. A. Meyers. 2011. Disease transmission in territorial populations: the small-world network of Serengeti lions. Journal of the Royal Society Interface 8:776-786. 129

Creel, S., G. Spong, N. Creel, and 2001. Interspecific competition and the population biology of extinction-prone carnivores. Pages 35-59 in D. Macdonald, J. L. Gittleman, R. Wayne, and S. Funk, editors. Carnivore conservation. Cambridge University Press.

Crowl, T. A., T. O. Crist, R. R. Parmenter, G. Belovsky, and A. E. Lugo. 2008. The spread of invasive species and infectious disease as drivers of ecosystem change. Frontiers in Ecology and the Environment 6:238-246.

Cunningham, M. W., D. B. Shindle, A. B. Allison, S. P. Terrell, D. G. Mead, and M. Owen. 2009. Canine distemper epizootic in Everglades mink. J Wildl Dis 45 :1150-1157.

Dalla Villa, P., S. Kahn, L. Stuardo, L. Iannetti, A. Di Nardo, and J. Serpell. 2010. Free- roaming dog control among OIE-member countries. Preventive Veterinary Medicine 97 :58-63.

Daniels, T. J. and M. Bekoff. 1989. Spatial and temporal resource use by feral and abandoned dogs. Ethology 81 :300-312.

Daoust, P.-Y., S. R. McBurney, D. L. Godson, M. W. Van De Bildt, and A. D. Osterhaus. 2009. Canine distemper virus-assosiated encephalitis in free-living Lynx (Lynx canadensis) and Bobcats (Lynx rufus) of eastern Canada Journal of Wildlife Diseases 45 :611-624.

Daszak, P., A. A. Cunningham, and A. D. Hyatt. 2000. Emerging infectious diseases of wildlife--threats to biodiversity and human health. Science 287 :443-449.

Davlin, S. L. and H. M. VonVille. 2012. Canine rabies vaccination and domestic dog population characteristics in the developing world: A systematic review. Vaccine 30 :3492-3502.

De Bosschere, H., S. Roels, N. Lemmens, and E. Vanopdenbosch. 2005. Canine distemper virus in Asian clawless otter (Aonyx cinereus) littermates in captivity. Vlaams Diergeneeskundig Tijdschrift 74 :299.

De Castro, F. and B. Bolker. 2005. Mechanisms of disease induced extinction. Ecology Letters 8:117-126.

130

Deem, S. L., L. H. Spelman, R. A. Yates, and R. J. Montali. 2000. Canine distemper in terrestrial carnivores: A review. Journal of Zoo and Wildlife Medicine 31 :441- 451.

Delgado, C. 2010. Plan de Manejo de la Reserva Costera Valdiviana. The Nature Conservancy.

Depue, J. E. and M. Ben David. 2010. River Otter Latrine Site Selection in Arid Habitats of Western Colorado, USA. The Journal of Wildlife Management 74 :1763-1767.

Diamond, J. M. and D. Ordunio. 1997. Guns, germs, and steel. Norton New York.

Dietz, R., J. Teilmann, S. M. Andersen, F. Rigét, and M. T. Olsen. 2013. Movements and site fidelity of harbour seals (Phoca vitulina) in Kattegat, Denmark, with implications for the epidemiology of the phocine distemper virus. ICES Journal of Marine Science: Journal du Conseil 70 :186-195.

Dobson, A. and J. Foufopoulos. 2001. Emerging infectious pathogens of wildlife. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 356 :1001-1012.

Domingo, M., L. Ferrer, M. Pumarola, A. Marco, J. Plana, S. Kennedy, M. McAliskey, and B. Rima. 1990. Morbillivirus in dolphins. Nature 348 :21.

Donadio, E. and S. W. Buskirk. 2006. Diet, morphology, and interspecific killing in carnivora. American Naturalist 167 :524-536.

Dovie, D. B., C. M. Shackleton, and E. Witkowski. 2006. Valuation of communal area livestock benefits, rural livelihoods and related policy issues. Land Use Policy 23 :260-271.

Dunkin, G. and P. Laidlaw. 1926. Studies in dog-distemper: I.—Dog-distemper in the ferret. Journal of Comparative Pathology and Therapeutics 39 :201-212.

Dunstone, N. 1993. The Mink. T. and A.D. Poyser Ltd, London, UK.

Dunstone, N., L. Durbin, I. Wyllie, R. Freer, G. A. Jamett, M. Mazzolli, and S. Rose. 2002. Spatial organization, ranging behaviour and habitat use of the kodkod (Oncifelis guigna) in southern Chile. Journal of Zoology 257 :1-11.

131

Elia, G., N. Decaro, V. Martella, F. Cirone, M. S. Lucente, E. Lorusso, L. Di Trani, and C. Buonavoglia. 2006. Detection of canine distemper virus in dogs by real-time RT-PCR. Journal of virological methods 136 :171-176.

Ernst, S., F. Metayer, and A. Huber. 1987. Influence of climatic factors in the variability of the prevalence of some canine infectious diseases. Archivos De Medicina Veterinaria 19 .

Fasola, L., C. Chehebar, D. W. Macdonald, G. Porro, and M. H. Cassini. 2009. Do alien North American mink compete for resources with native South American river otter in Argentinean Patagonia? Journal of Zoology 277 :187-195.

Fiorello, C. V., A. J. Noss, and S. L. Deem. 2006. Demography, hunting ecology, and pathogen exposure of domestic dogs in the Isoso of Bolivia. Conservation Biology 20 :762-771.

Flores-Ibarra, M. and G. Estrella-Valenzuela. 2004. Canine ecology and socioeconomic factors associated with dogs unvaccinated against rabies in a Mexican city across the US–Mexico border. Preventive Veterinary Medicine 62 :79-87.

Foreyt, W. J. and D. A. Jessup. 1982. Fatal pneumonia of bighorn sheep following association with domestic sheep. Journal of Wildlife Diseases 18 :163-168.

Fortin, D., H. L. Beyer, M. S. Boyce, D. W. Smith, T. Duchesne, and J. S. Mao. 2005. Wolves influence elk movements: behavior shapes a trophic cascade in Yellowstone National Park. Ecology 86 :1320-1330.

Fournier-Chambrillon, C., B. Aasted, A. Perrot, D. Pontier, F. Sauvage, M. Artois, J. M. Cassiede, X. Chauby, A. Dal Molin, C. Simon, and P. Fournier. 2004. Antibodies to aleutian mink disease parvovirus in free-ranging European mink (Mustela lutreola) and other small carnivores from Southwestern France. Journal of Wildlife Diseases 40 :394-402.

Funk, S., C. Fiorello, S. Cleaveland, and M. Gompper. 2001. The role of disease in carnivore ecology and conservation. Pages 443-466 in J. L. Gittleman, S. Funk, D. Macdonald, and R. Wayne, editors. Carnivore Conservation. Cambridge University Press,, Cambridge, UK,.

Gehring, T. M., K. C. VerCauteren, and J. M. Landry. 2010a. Livestock protection dogs in the 21st century: is an ancient tool relevant to modern conservation challenges? BioScience 60 :299-308.

132

Gehring, T. M., K. C. VerCauteren, M. L. Provost, and A. C. Cellar. 2010b. Utility of livestock-protection dogs for deterring wildlife from cattle farms. Wildlife Research 37 :715-721.

Geisel, O. 1979. Distemper in otters (Lutra lutra). Berliner Und Munchener Tierarztliche Wochenschrift 92 .

Gibbs, E., W. Taylor, M. J. Lawman, and J. Bryant. 2008. Classification of peste des petits ruminants virus as the fourth member of the genus Morbillivirus. Intervirology 11 :268-274.

Gittleman, J., S. Funk, D. Macdonald, and R. Wayne. 2001. Strategies for carnivore conservation: lessons from contemporary extinctions. Carnivore conservation 5:61-92.

Gittleman, J. L. 2001. Carnivore conservation. Cambridge University Press.

González, A., A. Novaro, M. Funes, O. Pailacura, M. J. Bolgeri, and S. Walker. 2012. Mixed-breed guarding dogs reduce conflict between goat herders and native carnivores in Patagonia. Human Wildlife Interactions 6:134-141.

Gonzalez-Acuña, D., R. Ortega-Vasquez, P. Rivera-Ramírez, and J. Cabello-Cabalin. 2003. Verdacht auf Staupe beim Graufuchs (Pseudalopex griseus) im mittleren Chile (Fallbericht). Zeitschrift fur Jagdwissenschaft:1-4.

Gorham, J. R. 1999. Some experiments and field observations of distemper in mink and ferrets. Advances in veterinary medicine 41 :557-570.

Gorman, M. L. and B. J. Trowbridge. 1989. The role of odor in the social lives of carnivores. Pages 57-88 Carnivore behavior, ecology, and evolution. Springer.

Goulson, D., S. A. Hawson, and J. C. Stout. 1998. Foraging bumblebees avoid flowers already visited by conspecifics or by other bumblebee species. Animal Behaviour 55 :199-206.

Gowtage-Sequeira, S., A. C. Banyard, T. Barrett, H. Buczkowski, S. M. Funk, and S. Cleaveland. 2009. Epidemiology, pathology, and genetic analysis of a canine distemper epidemic in Namibia. Journal of Wildlife Diseases 45 :1008-1020.

133

Grachev, M., V. Kumarev, L. Mamaev, V. Zorin, L. Baranova, N. Denikina, S. Belikov, E. Petrov, V. Kolesnik, and R. Kolesnik. 1989. Distemper virus in Baikal seals. Nature 338 .

Greene, C. E. 2006. Infectious diseases of the dog and cat. Saunders/ Elsevier Science.

Greene, C. E. and M. J. Appel. 2006. Canine distemper. Pages 25-41 in C. E. Greene, editor. Infectious diseases of the dog and cat. Saunders/Elsevier, Philadelphia ; London.

Greene, C. E. and M. Vandevelde. 2012. Canine Distemper. Pages 25-42 in C. E. Greene, editor. Infectious Diseases of the Dog and Cat. Elsevier, St Louis.

Grenfell, B. T. and A. P. Dobson. 1995. Ecology of infectious diseases in natural populations. Cambridge University Press.

Grinnell, J., C. Packer, and A. E. Pusey. 1995. Cooperation in male lions: kinship, reciprocity or mutualism? Animal Behaviour 49 :95-105.

Haas, L. and T. Barrett. 1996. Rinderpest and other animal morbillivirus infections: comparative aspects and recent developments. Journal of Veterinary Medicine, Series B 43 :411-420.

Harder, T. C. and A. D. Osterhaus. 1997. Canine distemper virus—a morbillivirus in search of new hosts? Trends in Microbiology 5:120-124.

Hart, B. L. 1995. Analysing breed and gender differences in behaviour. The Domestic Dog: its evolution, behaviour and interactions with people. Cambridge University Press, Cambridge:65-77.

Haydon, D. T., M. K. Laurenson, and C. Sillero Zubiri. 2002. Integrating epidemiology into population viability analysis: Managing the risk posed by rabies and canine distemper to the Ethiopian wolf. Conservation Biology 16 :1372-1385.

Hess, G. 1996. Disease in metapopulation models: implications for conservation. Ecology 77 :1617-1632.

Hosmer, D. W., T. Hosmer, S. Le Cessie, and S. Lemeshow. 1997. A comparison of goodness-of-fit tests for the logistic regression model. Statistics in medicine 16 :965-980.

134

Hughes, J. and D. W. Macdonald. 2013. A review of the interactions between free- roaming domestic dogs and wildlife. Biological Conservation 157 :341-351.

Hutchings, M. R. and P. C. L. White. 2000. Mustelid scent-marking in managed ecosystems: Implications for population management. Mammal Review 30 :157- 169.

INE. 2011. Encuesta de ganado ovino. in Instituto Nacional de Estadísticas, editor., Santiago,Chile.

IUCN. 2013. Red List of Threatened Species.Version 2012.2.

Jachowski, D. S., R. A. Gitzen, M. B. Grenier, B. Holmes, and J. J. Millspaugh. 2011. The importance of thinking big: Large-scale prey conservation drives black- footed ferret reintroduction success. Biological Conservation 144 :1560-1566.

Jaksic, F. M. 1998. Vertebrate invaders and their ecological impacts in Chile. Biodiversity & Conservation 7:1427-1445.

Jaksic, F. M., J. A. Iriarte, J. E. Jimenez, and D. R. Martinez. 2002. Invaders without frontiers: Cross-border invasions of exotic mammals. Biological Invasions 4:157- 173.

Jetz, W., D. S. Wilcove, and A. P. Dobson. 2007. Projected impacts of climate and land- use change on the global diversity of birds. PLoS Biology 5:e157.

Jiménez, J. 2007. Ecology of a coastal population of the critically endangered Darwin's fox (Pseudalopex fulvipes) on Chiloé Island, southern Chile. Journal of Zoology 271 :63-77.

Jiménez, J., C. Briceño, H. Alcaíno, P. Vásquez, S. Funk, and D. González-Acuña. 2012. Coprologic survey of endoparasites from Darwin’s fox (Pseudalopex fulvipes) in Chiloé, Chile. Archivos De Medicina Veterinaria 44 :93-97.

Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61 :65-71.

Joppa, L. N. and A. Pfaff. 2009. High and far: biases in the location of protected areas. Plos One 4:e8273.

135

Kameo, Y., Y. Nagao, Y. Nishio, H. Shimoda, H. Nakano, K. Suzuki, Y. Une, H. Sato, M. Shimojima, and K. Maeda. 2012. Epizootic canine distemper virus infection among wild mammals. Veterinary Microbiology 154 :222-229.

Kapil, S. and T. J. Yeary. 2011. Canine distemper spillover in domestic dogs from urban wildlife. The Veterinary clinics of North America. Small animal practice 41 :1069.

Kimber, K. R., G. V. Kollias, and E. J. Dubovi. 2000. Serologic survey of selected viral agents in recently captured wild North American river otters (Lontra canadensis). Journal of Zoo and Wildlife Medicine 31 :168-175.

Kleiman, D. 1966. Scent marking in the Canidae. Pages 167-177 in Symp. Zool. Soc. Lond.

Koster, J. 2008. The impact of hunting with dogs on wildlife harvests in the Bosawas Reserve, Nicaragua. Environmental Conservation 35 :211-220.

Kruuk, H. and H. Snell. 1981. Prey selection by feral dogs from a population of marine iguanas (Amblyrhynchus cristatus). Journal of Applied Ecology 18 :197-204.

Lacerda, A. C. R., W. M. Tomas, and J. Marinho. 2009. Domestic dogs as an edge effect in the Brasilia National Park, Brazil: interactions with native mammals. Animal Conservation 12 :477-487.

Lafferty, K. D. and L. R. Gerber. 2002. Good medicine for conservation biology: the intersection of epidemiology and conservation theory. Conservation Biology 16 :593-604.

Lambin, E. F., B. L. Turner, H. J. Geist, S. B. Agbola, A. Angelsen, J. W. Bruce, O. T. Coomes, R. Dirzo, G. Fischer, and C. Folke. 2001. The causes of land-use and land-cover change: moving beyond the myths. Global environmental change 11 :261-269.

Lembo, T., K. Hampson, D. T. Haydon, M. Craft, A. Dobson, J. Dushoff, E. Ernest, R. Hoare, M. Kaare, and T. Mlengeya. 2008. Exploring reservoir dynamics: a case study of rabies in the Serengeti ecosystem. Journal of Applied Ecology 45 :1246- 1257.

Loomis, J., P. Kent, L. Strange, K. Fausch, and A. Covich. 2000. Measuring the total economic value of restoring ecosystem services in an impaired river basin: results from a contingent valuation survey. Ecological Economics 33 :103-117.

136

López, J., K. Abarca, J. Cerda, B. Valenzuela, L. Lorca, A. Olea, and X. Aguilera. 2009. Surveillance system for infectious diseases of pets, Santiago, Chile. Emerging Infectious Diseases 15 :1674.

Luebert, F. and P. Pliscoff. 2005. Bioclimas de la Cordillera de la Costa del Centro-sur de ChileBioclimas de la Cordillera de la Costa del Centro-sur de Chile. Pages 61-74 in C. Smith-Ramirez, J. Armesto, and C. Valdovinos, editors. Historia, biodiversidad y ecología de los bosques costeros de Chile. Editorial Universitaria, Santiago. Chile.

Lyles, A. M. and A. P. Dobson. 1993. Infectious disease and intensive management: population dynamics, threatened hosts, and their parasites. Journal of Zoo and Wildlife Medicine 24 :315-326.

MacDonald, D. W. 1980. Patterns of scent marking with urine and faeces amongst carnivore communities. Pages 107-139 in Symposia of the Zoological Society of London.

MacFarland, C. G., J. Villa, and B. Toro. 1974. The Galápagos giant tortoises (Geochelone elephantopus) Part I: Status of the surviving populations. Biological Conservation 6:118-133.

Madsen, A., H. Dietz, P. Henriksen, and B. Clausen. 1999. Survey of Danish Free Living Otters Lutra lutra-a Consecutive Collection and Necroscopy of Dead Bodies IUCN Otter Spec. Group Bull 16 :65-76.

Maestas, J. D., R. L. Knight, and W. C. Gilgert. 2003. Biodiversity across a Rural Land- Use Gradient. Conservation Biology 17 :1425-1434.

Mamaev, L., N. Denikina, S. Belikov, V. Volchkov, I. Visser, M. Fleming, C. Kai, T. Harder, B. Liess, and A. Osterhaus. 1995. Characterisation of morbilliviruses isolated from Lake Baikal seals (Phoca sibirica). Veterinary Microbiology 44 :251-259.

Matter, H. C., A. I. Wandeler, B. E. Neuenschwander, L. Harischandra, and F. X. Meslin. 2000. Study of the dog population and the rabies control activities in the Mirigama area of Sri Lanka. Acta Tropica 75 :95-108.

May, R. M. 1986. Species conservation: The cautionary tale of the black-footed ferret. Nature 320 :13-14.

137

McCallum, H. and A. Dobson. 1995. Detecting disease and parasite threats to endangered species and ecosystems. Trends in Ecology & Evolution 10 :190-194.

McCullough, S., F. McNeilly, G. Allan, S. Kennedy, J. Smyth, S. Cosby, S. McQuaid, and B. Rima. 1991. Isolation and characterisation of a porpoise morbillivirus. Archives of Virology 118 :247-252.

McKinney, M. L. 2002. Urbanization, Biodiversity, and Conservation: The impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. BioScience 52 :883-890.

Medina G. 1996. Conservation and status of Lutra provocax in Chile. Pacific Conservation Biology 2: 414-419.

Medina, G. 1997. A comparison of the diet and distribution of southern river otter (Lutra provocax) and mink (Mustela vison) in southern Chile. Journal of Zoology 242 :291-297.

Medina-Vogel, G., Kaufman, V.S., Monsalve, R., Gomez, V. 2003. The influence of riparian vegetation, woody debris, stream morphology and human activity on the use of rivers by southern river otters in Lontra provocax in Chile. Oryx 37 : 422- 430.

Medina Vogel, G., M. Barros, J. Organ, and L. Bonesi. 2013. Coexistence between the southern river otter and the alien invasive North American mink in marine habitats of southern Chile. Journal of Zoology 290 :27-34.

Meek, P. D. 1999. The movement, roaming behaviour and home range of free-roaming domestic dogs, Canis lupus familiaris, in coastal New South Wales. Wildlife Research 26 :847-855.

Meli, M. L., P. Simmler, V. Cattori, F. Martínez, A. Vargas, F. Palomares, J. V. López- Bao, M. A. Simón, G. López, and L. León-Vizcaino. 2010. Importance of canine distemper virus (CDV) infection in free-ranging Iberian lynxes (Lynx pardinus). Veterinary Microbiology 146 :132-137.

Miller, M., I. Gardner, C. Kreuder, D. Paradies, K. Worcester, D. Jessup, E. Dodd, M. Harris, J. Ames, and A. Packham. 2002. Coastal freshwater runoff is a risk factor for Toxoplasma gondii infection of southern sea otters (Enhydra lutris nereis). International Journal for Parasitology 32 :997-1006.

138

Morgan, E., M. Lundervold, G. Medley, B. Shaikenov, P. Torgerson, and E. Milner- Gulland. 2006. Assessing risks of disease transmission between wildlife and livestock: the Saiga antelope as a case study. Biological Conservation 131 :244- 254.

Mos, L., P. Ross, D. McIntosh, and S. Raverty. 2003. Canine distemper virus in river otters in British Columbia as an emergent risk for coastal pinnipeds. Veterinary Record 152 :237-239.

Munson, L. 1994. Canine distemper epidemic in Serengeti lions. Small Carnivore Conservation 11 :7.

Munson, L., K. A. Terio, R. Kock, T. Mlengeya, M. E. Roelke, E. Dubovi, B. Summers, A. R. Sinclair, and C. Packer. 2008. Climate extremes promote fatal co-infections during canine distemper epidemics in African lions. Plos One 3:e2545.

Murray, D. L., C. A. Kapke, J. F. Evermann, and T. K. Fuller. 1999. Infectious disease and the conservation of free ranging large carnivores. Animal Conservation 2:241-254.

Myers, D. L., A. Zurbriggen, H. Lutz, and A. Pospischil. 1997. Distemper: not a new disease in lions and tigers. Clinical and Diagnostic Laboratory Immunology 4:180-184.

Naughton-Treves, L., M. B. Holland, and K. Brandon. 2005. The role of protected areas in conserving biodiversity and sustaining local livelihoods. Annu. Rev. Environ. Resour. 30 :219-252.

Nava, A. F. D., L. Cullen Jr, D. A. Sana, M. S. Nardi, J. D. Ramos Filho, T. F. Lima, K. C. Abreu, and F. Ferreira. 2008. First evidence of canine distemper in Brazilian free-ranging felids. Ecohealth 5:513-518.

Nelson, B., M. Hebblewhite, V. Ezenwa, T. Shury, E. H. Merrill, P. C. Paquet, F. Schmiegelow, D. Seip, G. Skinner, and N. Webb. 2012. Prevalence of antibodies to canine parvovirus and distemper virus in wolves in the Canadian Rocky Mountains. Journal of Wildlife Diseases 48 :68.

Newbury, S., L. J. Larson, and R. D. Schultz. 2009. Canine distemper virus. Infectious Disease Management in Animal Shelters. Ames, IA: Wileyand Blackwell:161- 173.

139

Olson, D. M. and E. Dinerstein. 1998. The global 200: A representation approach to conserving the Earth's most biologically valuable ecoregions. Conservation Biology 12 :502-515.

Oni, O., W. Wajjwalku, O. Boodde, and W. Chumsing. 2006. Canine distemper virus antibodies in the Asian elephant (Elaphas maximus). Veterinary Record 159 :420- 421.

Ordeñana, M. A., K. R. Crooks, E. E. Boydston, R. N. Fisher, L. M. Lyren, S. Siudyla, C. D. Haas, S. Harris, S. A. Hathaway, and G. M. Turschak. 2010. Effects of urbanization on carnivore species distribution and richness. Journal of Mammalogy 91 :1322-1331.

Origgi, F., U. Sattler, P. Pilo, and A. Waldvogel. 2013. Fatal Combined Infection With Canine Distemper Virus and Orthopoxvirus in a Group of Asian Marmots (Marmota caudata). Veterinary Pathology.

Osterhaus, A., J. Groen, H. Spijkers, H. Broeders, F. UytdeHaag, P. De Vries, J. Teppema, I. Visser, M. Van de Bildt, and E. Vedder. 1990. Mass mortality in seals caused by a newly discovered virus-like morbillivirus. Veterinary Microbiology 23 :343-350.

Page, L. K., R. K. Swihart, and K. R. Kazacos. 1999. Implications of raccoon latrines in the epizootiology of baylisascariasis. Journal of Wildlife Diseases 35 :474-480.

Palomares, F. and T. M. Caro. 1999. Interspecific killing among mammalian carnivores. The American naturalist 153 :492-508.

Panzera, Y., M. G. Calderón, N. Sarute, S. Guasco, A. Cardeillac, B. Bonilla, M. Hernández, L. Francia, G. Bedó, and J. La Torre. 2012. Evidence of two co- circulating genetic lineages of canine distemper virus in South America. Virus Research 163 :401-404.

Paquet, P. C. 1992. Prey use strategies of sympatric wolves and coyotes in Riding Mountain National Park, Manitoba. Journal of Mammalogy 73 :337-343.

Perdrau, J. and L. Pugh. 1930. The pathology of disseminated encephalomyelitis of the dog (the "nervous form of canine distemper"). The Journal of Pathology and Bacteriology 33 :79-91.

140

Pita, R., A. Mira, F. Moreira, R. Morgado, and P. Beja. 2009. Influence of landscape characteristics on carnivore diversity and abundance in Mediterranean farmland. Agriculture, Ecosystems and Environment 132 :57-65.

Plowright, R. K., P. Foley, H. E. Field, A. P. Dobson, J. E. Foley, P. Eby, and P. Daszak. 2011. Urban habituation, ecological connectivity and epidemic dampening: the emergence of Hendra virus from flying foxes (Pteropus spp.). Proceedings of the Royal Society B: Biological Sciences 278 :3703-3712.

Prager, K., J. A. Mazet, E. J. Dubovi, L. G. Frank, L. Munson, A. P. Wagner, and R. Woodroffe. 2013. Rabies Virus and Canine Distemper Virus in Wild and Domestic Carnivores in Northern Kenya: Are Domestic Dogs the Reservoir? Ecohealth 9:1-16.

R Development Core Team. 2011. R: A language and environment for statistical computing. . R Foundation for Statistical Computing, Vienna, Austria.

Ratsitorahina, M., J. H. Rasambainarivo, S. Raharimanana, H. Rakotonandrasana, M. P. Andriamiarisoa, F. A. Rakalomanana, and V. Richard. 2009. Dog ecology and demography in Antananarivo, 2007. BMC veterinary research 5:21.

Reece, J. 2005. Dogs and dog control in developing countries. The State of the Animals III:55-64.

Rich, L. N., M. S. Mitchell, J. A. Gude, and C. A. Sime. 2012. Anthropogenic mortality, intraspecific competition, and prey availability influence territory sizes of wolves in Montana. Journal of Mammalogy 93 :722-731.

Rigg, R. 2001. Livestock guarding dogs: their current use world wide. in I. S. C. S. Group, editor. Canid Specialist Group.

Rodríguez-Jorquera, I. and M. A. Sepúlveda. 2011. Trophic spatial variations in the Southern river otter, Lontra provocax, in freshwater habitats, Chile. Pages 70 - 75 in Proceedings of XIth International Otter Colloquium. IUCN Otter Spec. Group Bull.

Roelke-Parker, M. E., L. Munson, C. Packer, R. Kock, S. Cleaveland, M. Carpenter, S. J. O'Brien, A. Pospischil, R. Hofmann-Lehmann, and H. Lutz. 1996. A canine distemper virus epidemic in Serengeti lions (Panthera leo). Nature 379 :441-445.

141

Roemer, G. W., C. J. Donlan, and F. Courchamp. 2002. Golden eagles, feral pigs, and insular carnivores: how exotic species turn native predators into prey. Proceedings of the National Academy of Sciences 99 :791-796.

Saito, T. B., A. A. Alﬁeri, S. R. Wosiacki, F. J. Negrao, H. S. A. Morais, and A. F. Alﬁeri. 2006. Detection of canine distemper virus by reverse transcriptase- polymerase chain reaction in the urine of dogs with clinical signs of distemper encephalitis. Research Veterinary Science 80 :116–119.

Sakai, K., N. Nagata, Y. Ami, F. Seki, Y. Suzaki, N. Iwata-Yoshikawa, T. Suzuki, S. Fukushi, T. Mizutani, and T. Yoshikawa. 2013. Lethal Canine Distemper Virus Outbreak in Cynomolgus Monkeys in Japan in 2008. Journal of Virology 87 :1105-1114.

Schaefer, J. M., R. D. Andrews, and J. J. Dinsmore. 1981. An assessment of coyote and dog predation on sheep in southern Iowa. The Journal of Wildlife Management 45 :883-893.

Schmitt, S. M., D. J. O'brien, C. S. Bruning-Fann, and S. D. Fitzgerald. 2002. Bovine tuberculosis in Michigan wildlife and livestock. Annals of the New York Academy of Sciences 969 :262-268.

Schulte, B. A. 1998. Scent marking and responses to male castor fluid by beavers. Journal of Mammalogy 79 :191-203.

Scott, M. D. and K. Causey. 1973. Ecology of feral dogs in Alabama. Journal of Wildlife Management 37 :253-265.

Seaman, D. E. and R. A. Powell. 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology 77 :2075-2085.

Sepúlveda, M., J. L. Bartheld, R. Monsalve, V. Gómez, and G. Medina-Vogel. 2007. Habitat use and spatial behaviour of the endangered Southern river otter (Lontra provocax) in riparian habitats of Chile: conservation implications. Biological Conservation 140:329-338.

Sepúlveda, M., Franco, M., Medina, G., Fasola, L., Alvarez, R. 2008. Lontra provocax. In: IUCN 2012. IUCN Red List of Threatened Species.

142

Sepúlveda, M., J. Bartheld, C. Meynard, M. Benavides, C. Astorga, D. Parra, and G. Medina Vogel. 2009. Landscape features and crustacean prey as predictors of the Southern river otter distribution in Chile. Animal Conservation 12 :522-530.

Sepúlveda, M. A., C. Munoz-Zanzi, C. Rosenfeld, R. Jara, K. M. Pelican, and D. Hill. 2011. Toxoplasma gondii in feral American minks at the Maullin river, Chile. Veterinary Parasitology 175 :60-65.

Shackleton, S., C. Shackleton, T. Netshiluvhi, B. Geach, A. Ballance, and D. Fairbanks. 2002. Use patterns and value of savanna resources in three rural villages in South Africa. Economic Botany 56 :130-146.

Shen, D. and J. Gorham. 1978. Contact transmission of distemper virus in ferrets. Research in Veterinary Science 24 :118.

Shen, D. and J. Gorham. 1980. Survival of pathogenic distemper virus at 5C and 25C. Veterinary Medicine & Small Animal Clinician 75 :69-70.

Sillero Zubiri, C., D. W. Macdonald, and A. A. King. 1996. Rabies and mortality in ethiopian wolves (Canis simensis). Journal of Wildlife Diseases 32 :80-86.

Silva-Rodriguez, E. A., G. R. Ortega-Solis, and J. E. Jimenez. 2010. Conservation and ecological implications of the use of space by chilla foxes and free-ranging dogs in a human-dominated landscape in southern Chile. Austral Ecology 35 :765-777.

Silva-Rodriguez, E. A. and K. E. Sieving. 2011. Influence of care of domestic carnivores on their predation on vertebrates. Conservation Biology 25 :808-815.

Silva-Rodríguez, E. A. and K. E. Sieving. 2012. Domestic dogs shape the landscape-scale distribution of a threatened forest ungulate. Biological Conservation 150 :103-110.

Simberloff, D. 2005. Non-native species do threaten the natural environment! Journal of Agricultural and Environmental Ethics 18 :595-607.

Simon-Martinez, J., R. Ulloa-Arvizu, V. Soriano, and R. Fajardo. 2008. Identification of a genetic variant of canine distemper virus from clinical cases in two vaccinated dogs in Mexico. The Veterinary Journal 175 :423-426.

Simpson, V. R. 2006. Patterns and significance of bite wounds in Eurasian otters (Lutra Lutra) in southern and south-west England. Veterinary Record 158 :113-119.

143

Skerratt, L. F., L. Berger, R. Speare, S. Cashins, K. R. McDonald, A. D. Phillott, H. B. Hines, and N. Kenyon. 2007. Spread of chytridiomycosis has caused the rapid global decline and extinction of frogs. Ecohealth 4:125-134.

Steinfeld, H., P. Gerber, T. Wassenaar, V. Castel, and C. De Haan. 2006. Livestock's long shadow: environmental issues and options. Food & Agriculture Org.

Stowhas, P. 2012. Análisis del conflicto entre carnívoros silvestres y campesinos en el sur de Chile. Universidad Mayor, Santiago de Chile.

Sun, Z., A. Li, H. Ye, Y. Shi, Z. Hu, and L. Zeng. 2010. Natural infection with canine distemper virus in hand-feeding Rhesus monkeys in China. Veterinary Microbiology 141 :374-378.

Sundqvist, A. K., H. Ellegren, and C. Vilà. 2008. Wolf or dog? Genetic identification of predators from saliva collected around bite wounds on prey. Conservation Genetics 9:1275-1279.

Svartberg, K. 2002. Shyness–boldness predicts performance in working dogs. Applied Animal Behaviour Science 79 :157-174.

Swango, L. 1995. Canine viral diseases. Textbook of Veterinary Internal Medicine, ed 3:298-311.

Taborsky, M. 1988. Kiwis and dog predation: observations in Waitangi State Forest. Notornis 35 :197-202.

Tennant, B., R. Gaskell, R. Jones, and C. Gaskell. 1991. Prevalence of antibodies to four major canine viral diseases in dogs in a Liverpool hospital population. Journal of Small Animal Practice 32 :175-179.

Thorne, E. and E. S. Williams. 1988. Disease and Endangered Species: The Black footed Ferret as a Recent Example. Conservation Biology 2:66-74.

Tompkins, D., A. White, and M. Boots. 2003. Ecological replacement of native red squirrels by invasive greys driven by disease. Ecology Letters 6:189-196.

Tuyttens, F., R. Delahay, D. Macdonald, C. Cheeseman, B. Long, and C. Donnelly. 2000. Spatial perturbation caused by a badger (Meles meles) culling operation: implications for the function of territoriality and the control of bovine tuberculosis (Mycobacterium bovis). Journal of Animal Ecology 69 :815-828. 144

Valenzuela, A. E., A. R. Rey, L. Fasola, and A. Schiavini. 2013. Understanding the interspecific dynamics of two co-existing predators in the Tierra del Fuego Archipelago: the native southern river otter and the exotic American mink. Biological Invasions 15 :1-12.

Vanak, A. T. and M. E. Gompper. 2009. Dogs Canis familiaris as carnivores: their role and function in intraguild competition. Mammal Review 39 :265–283.

Vanak, A. T. and M. E. Gompper. 2010. Interference competition at the landscape level: the effect of free-ranging dogs on a native mesocarnivore. Journal of Applied Ecology 47 :1225-1232.

Veblen, T. T., M. Mermoz, C. Martin, and T. Kitzberger. 2003. Ecological impacts of introduced animals in Nahuel Huapi National Park, Argentina. Conservation Biology 6:71-83.

Vila, C., J. E. Maldonado, and R. K. Wayne. 1999. Phylogenetic relationships, evolution, and genetic diversity of the domestic dog. Journal of Heredity 90 :71-77.

Vitousek, P. M., C. M. D’Antonio, L. L. Loope, and R. Westbrooks. 1996. Biological invasions as global environmental change. American Scientist 84 :468-478. von Messling, V., N. Svitek, and R. Cattaneo. 2006. Receptor (SLAM [CD150]) recognition and the V protein sustain swift lymphocyte-based invasion of mucosal tissue and lymphatic organs by a morbillivirus. Journal of Virology 80 :6084- 6092.

Waters-Bayer, A. and W. Bayer. 1992. The role of livestock in the rural economy. Nomadic Peoples 31 :3-18.

Weldon, C., L. H. du Preez, A. D. Hyatt, R. Muller, and R. Speare. 2004. Origin of the amphibian chytrid fungus. Emerging Infectious Diseases 10 :2100.

White, P. C. L., M. Böhm, G. Marion, and M. R. Hutchings. 2008. Control of bovine tuberculosis in British livestock: there is no ‘silver bullet’. Trends in Microbiology 16 :420-427.

Williams, E. S., E. T. Thorne, M. J. Appel, and D. W. Belitsky. 1988. Canine distemper in black-footed ferrets (Mustela nigripes) from Wyoming. Journal of Wildlife Diseases 24 :385-398.

145

Wittemyer, G., P. Elsen, W. T. Bean, A. C. O. Burton, and J. S. Brashares. 2008. Accelerated human population growth at protected area edges. Science 321 :123- 126.

Wolfe, N. D., P. Daszak, A. M. Kilpatrick, and D. S. Burke. 2005. Bushmeat hunting, deforestation, and prediction of zoonotic disease. Emerging Infectious Diseases 11 :1822.

Wolfe, N. D., C. P. Dunavan, and J. Diamond. 2007. Origins of major human infectious diseases. Nature 447 :279-283.

Woma, T. Y., M. van Vuuren, A.-M. Bosman, M. Quan, and M. Oosthuizen. 2010. Phylogenetic analysis of the haemagglutinin gene of current wild-type canine distemper viruses from South Africa: Lineage Africa. Veterinary Microbiology 143 :126-132.

Woo, P. C., S. K. Lau, B. H. Wong, R. Y. Fan, A. Y. Wong, A. J. Zhang, Y. Wu, G. K. Choi, K. S. Li, and J. Hui. 2012. Feline morbillivirus, a previously undescribed paramyxovirus associated with tubulointerstitial nephritis in domestic cats. Proceedings of the National Academy of Sciences 109 :5435-5440.

Woodroffe, R. 1999. Managing disease threats to wild mammals. Animal Conservation 2:185-193.

Woodroffe, R. and C. A. Donnelly. 2011. Risk of contact between endangered African wild dogs Lycaon pictus and domestic dogs: opportunities for pathogen transmission. Journal of Applied Ecology.

Woolhouse, M. E., S. Gowtage-Sequeria, M. Woolhouse, and S. Gowtage-Sequeria. 2005. Host range and emerging and reemerging pathogens. Emerging Infectious Diseases 11 :1842.

Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home- range studies. Ecology 70 :164-168.

Yoshikawa, Y., F. Ochikubo, Y. Matsubara, H. Tsuruoka, M. Ishii, K. Shirota, Y. Nomura, M. Sugiyama, and K. Yamanouchi. 1989. Natural infection with canine distemper virus in a Japanese monkey (Macaca fuscata). Veterinary Microbiology 20 :193-205.

146

Young, J. K., K. A. Olson, R. P. Reading, S. Amgalanbaatar, and J. Berger. 2011. Is Wildlife Going to the Dogs? Impacts of Feral and Free-roaming Dogs on Wildlife Populations. BioScience 61 :125-132.

Zhao, J.-J., X.-J. Yan, X.-L. Chai, V. Martella, G.-L. Luo, H.-L. Zhang, H. Gao, Y.-X. Liu, X. Bai, and L. Zhang. 2010. Phylogenetic analysis of the haemagglutinin gene of canine distemper virus strains detected from breeding foxes, raccoon dogs and minks in China. Veterinary Microbiology 140 :34-42.

Zuberogoitia, I., J. Zabala, and J. A. Martinez. 2006. Diurnal activity and observations of the hunting and ranging behaviour of the American mink (Mustela vison). Mammalia 70 :310-312.

147