Tuesday, July 15, 2014

Guarding Sheep: Risks of an Ancient Function in the Modern (Ever More Crowded) World

Livestock Guarding Dog Fighting Coyote (USDA)
An attack of a bicyclist in a national forest in Colorado has been sent back to lower courts by the Colorado Supreme Court for a determination of whether the sheep guarding dogs that attacked the bicyclist were under the control of their owner at the time of the attack.  What control will mean in this case is uncertain because the appellate court had not seen control of the dogs as an issue but rather control of the land on which the attack occurred. 

Defining control with regard to sheep guarding dogs will not be easy.  Unlike suspect apprehension dogs, which are under the control of a police handler even if temporarily out of the handler’s sight, livestock guarding dogs spend most of their time protecting sheep, often with the handler miles away.  Indeed, their value is in reducing predation by being where the shepherd cannot be.  Some speculation on what control will mean here is now appropriate, and it is to be hoped that the case will continue, though the costs to the litigants, financial and emotional, may bring it to an end before such questions can be answered. 

First, however, let us look at the history of the case, and of livestock guarding in general.

An Attack by Sheep Guard Dogs in Colorado

Over 2.75 million sheep graze on 614 million acres of public lands in the United States, making this one of the largest uses of federal lands.  Most western ranchers depend heavily on federal or state trust lands for grazing and access to those lands is essential to their economy.  Samuel and Cheri Robinson, like many sheep owners in the U.S., used guarding dogs to protect their herds against mountain lions, wolves, coyotes, feral dogs, and other predators. 

Renee Legro was in a bicycle race in the White River National Forest when she was attacked by the two sheep guarding dogs owned by the Robinsons.  The Robinsons had a permit from the U.S. Forest Service to graze sheep in the area, and the Vail Recreation District had a permit to periodically conduct the bicycle races on roads in the Forest.  The attack resulted in serious injuries to Legro, who sued the Robinsons. 

Long-time readers of this blog may recognize the facts, which were described in a prior blog concerning a decision of a Colorado appellate court. The matter has now been considered by the Colorado Supreme Court.  Robinson v. Legro, No. 12SC1002, 2014 CO 40, 2014 Colo. LEXIS 414 (May 27, 2014).

The Colorado Supreme Court provides additional detail regarding the attack:

“Renee Legro was attacked on a public road by two of the Robinsons' dogs, Tiny and Pastor, while participating in a mountain bike race sponsored by the Vail Recreation District. The road is located on land that both the Robinsons and the Vail Recreation District were entitled, by permit, to access. Ms. Legro sustained serious injuries during the attack. Neither the Robinsons nor their employees were near the scene. The Robinsons' shepherd was about a little over a mile away from the area of the incident, ‘down by the river, trying to get the sheep to move along,’ at the time of the attack; he did not hear about the incident until the next day. Campers intervened to help Ms. Legro.”

Grazing Permit

The court then describes the sheep ranching operation of the Robinsons as follows:

“Samuel and Cheri Robinson are sheep ranchers who hold a 'Term Grazing Permit' issued by the United States Forest Service. The permit allows the Robinsons to graze a certain number of sheep on federal land within the White River National Forest in Eagle County for a period of ten years, at which time they must reapply for the permit. The Robinsons owned several Great Pyrenees dogs to protect their sheep from predators. Predator control dogs bond with and protect livestock by patrolling the grazing area, alerting the livestock to potential threats such as coyotes and bears, and chasing predators away as necessary.”

Colorado Dog Bite Statute

Colorado has a strict-liability dog bite statute which provides that a “person … who suffers serious bodily injury or death from being bitten by a dog while lawfully on public or private property shall be entitled to bring a civil action to recover economic damages against the dog owner regardless of the viciousness or dangerous propensities of the dog or the dog owner's knowledge or lack of knowledge of the dog's viciousness or dangerous propensities.” Colorado Revised Statutes (CRS) 13-21-124

There are six exemptions to strict liability, however, one of which is a “working dog” exemption that provides that a “dog owner shall not be liable to a person who suffers bodily injury, serious bodily injury, or death from being bitten by the dog: … (f) While the dog is working as a hunting dog, herding dog, farm or ranch dog, or predator control dog on the property of or under the control of the dog’s owner.”  CRS 13-21-124(5)(f).

Issue before Colorado Supreme Court

Guard Dogs with Sheep (USDA)
The Colorado court of appeals had interpreted the phrase, “on the property of or under the control of the dog’s owner,” as requiring that the dogs were working as predator control dogs on the dog owner’s property or on property under the dog owner’s control.  Thus, the question for that court was whether the bite occurred on property that was under their control as a result of the grazing permit.  Control was defined by that court to mean at least “sufficient control over the property such that a dog owner has the right to exclude persons from the property.” 

The question, according to the Colorado Supreme Court, was whether the appellate court had correctly interpreted the phrase, “on the property of or under the control of the dog’s owner,” in saying that the control question had to be resolved by a determination of whether the bite occurred on property under the Robinsons’ control.  The problem, the Supreme Court said, was almost grammatical:

“While we acknowledge that there is some facial ambiguity over whether it is the property or the dog that must be controlled by the dog owner, the more grammatically correct and logical reading of the exemption is that ‘on the property of’ and ‘under the control of’ modify ‘[w]hile the dog is working.’ A dog owner is therefore exempt from strict liability if a person is bitten by a predator control dog ‘while the dog is working’ either (a) ‘on the property of . . . the dog's owner’; or (b) ‘under the control of the dog's owner.’ Thus, it is control of the dog, not control of the property, that is the relevant inquiry.”

Other Types of Working Dogs

The Colorado Supreme Court notes that the working dog exemption to strict liability for a dog bite applies also to hunting dogs, herding dogs, farm or ranch dogs, as well as to predator control dogs such as the Robinsons used.  The court states:

“Much of these dogs' ‘work’ occurs on public lands; yet, the court of appeals would allow their owners to invoke the exemption only when they could lawfully exclude others from public lands--a right that these hunters, farmers, and ranchers rarely, if ever, have on federally owned lands--practically eviscerating the exemption.”

The court notes that the exemption for working dogs “limits strict liability for bites that occur while dogs are being used to complete socially beneficial tasks.” 

Court of Appeals Nevertheless Affirmed

Despite determining that the appellate court had posed the wrong question regarding the application of the strict liability exemption to the Robinsons, the Supreme Court affirmed the decision of the appellate court to return the case to the trial court.  That court will now be required to determine whether the dogs were under the control of the Robinsons at the time of the bite.  Since sheep guard dogs regularly work independently, living with a herd, not with the master such as sheep herding dogs do—indeed the concept of a handler is often irrelevant to the kind of training such dogs receive—the concept of “control” is not the same here as it is for most other trained working dogs. Consequently, the matter could be far from over.    

When does the owner have control of a sheep guarding dog? The shepherd was over a mile away at the time of the incident and did not know about it for a day.  Yet this is to be expected with sheep guarding dogs. They are particularly valuable when there is no one else protecting the sheep.  If shepherds were always nearby, they would often not be needed.  So does “control” require anticipating the presence of bicyclists who might come near the herd and training the dogs to accept the presence of humans? 


The USDA in its brochure, Livestock Guarding Dogs: Protecting Sheep from Predators, provides certain cautions regarding the use of livestock guarding dogs, including advising owners to do the following:
  • Alert neighbors that the dog may wander onto their property and enlist their aid in preventing roaming.
  • Post their property as to the presence of a dog.
  • Keep the dog off roads.
  • Be alert to the presence of poison baits, rodenticides, traps, and snares, and take appropriate precautions.
Although this is worded with the ownership of land in mind, it can be adapted to a situation involving a grazing permit.  Thus, sheep owners with such a permit should notify the authority granting the permit that they will be using livestock guarding dogs on the property to which they have been given access.  If some posting is possible, it should be done.  Keeping the dog off roads can perhaps best be accomplished by regularly checking the location of the herd. 

J.R. Lorenz and L. Coppinger, in their manual, Raising and Training a Livestock-guarding Dog, note that training goals depend in part on where a dog will be working.   Some dogs will have to work “in distant pastures, away from the house, and away from constant shepherding.”  They state: 

“[A] commercial producer with several hundred sheep may require a dog that is shy of people. A dog that prefers sheep to people will work better in unsupervised settings. Shyness to people can be fostered by minimizing human attention, beginning at 5 weeks of age.”

The issue of whether a working dog must be under a handler’s control has been addressed numerous times by courts in the context of serious and fatal attacks by police dogs.  (See Police and Military Dogs, Chapter 20: Suspect Apprehension and Bite Issues.)  In that context it has been determined that dogs released to find or apprehend a suspect can be out of the handler’s eyesight for brief periods if necessary because of the circumstances of the search, such as where a suspect is hiding in dark warehouse or a dense forest.  Such situations, however, are not analogous to the work of livestock guarding dogs, which are taught to identify with the sheep or other livestock they guard, and should often be kept out of the house so as not to become pets or to feel more comfortable with humans than with sheep.

On the other hand, having dogs that protect sheep against predators in areas where humans are also common should bring a degree of responsibility for the risks such dogs might bring to the humans.  Does that responsibility fit within the definition of control?  Perhaps, but certainly not with presumed handler proximity of a suspect apprehension or tracking dog.  How, then, is control to be defined?  By posing the question as one of control of the property, the appellate court at least found an approach that allowed for a determination of the meaning of control.  Following the Colorado Supreme Court’s logic will require the courts faced with defining control to explore areas where no easy precedent is available.  One area which will have to be considered involves nuisance cases where livestock guarding dogs are the nuisance. 

Nuisance Cases

Courts have exempted livestock guarding dogs from nuisance liability. In Hood River Country v. Mazzara, 193 Ore. App. 272, 89 P.3d 1195 (2004), six hours of barking would have ordinarily been a nuisance liability, and the complainant could have obtained relief if there were proof that the dogs were not barking at predators.  In the absence of such proof, however, the barking could have been a “farm practice” under Oregon statutory law.  The dogs were used to guard goats. 

“The county, for its part, introduced no evidence that use of livestock guardian dogs was not a farming practice or that defendant was not using her dogs in a generally accepted, reasonable, and prudent manner. The county focused on the fact that the dog barked for six hours, but it did not adduce any evidence that the dog was not in fact reacting to a predator or that barking of that duration could not, somehow, meet the definition of a farming practice.”

This was stated by an Oregon appellate court, reversing the trial court.  The trial judge had engaged in the following colloquy with the attorney for the goatherd:

"[DEFENSE COUNSEL]: If [the complaining neighbor is] unable to verify whether there was or was not a predator around, Mr. Tomson--or Mr. Johnson, from the Oregon Department of Agriculture testified these dogs sense a lot more than humans do. They'll sense by hearing better than humans, smell better than humans, they may know something's there when humans don't, and that's a completely appropriate response for that dog to bark to keep that predator away, and if it takes six hours, it takes six hours.

"THE COURT: Well, I think that's patently ridiculous, and I would never accept that. I'm not venting on you; I'm just venting on the idea that it would be appropriate for some dog out to just bark, bark, bark its life away out there, and that it is not reasonable for an owner of a dog, if it really wants that dog to protect their animals, to not be there to respond to it, to just expect this dog to be out there in an area where there's people that are going to be disturbed. I just don't accept that within the Agricultural Practice Act, and I'm not going to.

"So, if we're arguing about it's okay for six hours of barking, eight hours of barking, two hours of barking, I have the facts of this case and the facts here are that it looks like it's in excess of six hours of barking by this dog, for no apparent purpose that I have. We don't have any testimony other than supposition, and I'm not going to make my decision on supposition, that there may have been a predator somewhere in the world there. That's not going to fly.”

The appellate court concluded that the trial judge was wrong to feel as he had.  Perhaps the trial judge had never been near an outdoor kennel full of beagles trained to hunt when a coyote is in the neighborhood.    

To the same effect was a 2006 New York case, Groat v. Brennan, 831 NYS2d 353 (2006), where Great Pyrenees were used to guard alpacas.  The court stated:

“The record shows that Great Pyrenees dogs have been considered a major benefit to herding agriculture for over 1,000 years. The record also supports a finding that the immediate area of the Serino farm does have coyotes, and that while the Serinos did not lose any alpacas to predation prior to acquiring the dogs, alpacas are very susceptible to such attacks. It is further noted that the type of alpacas kept by the Serino respondents are valued at between $10,000 and $15,000 each. Thus, the loss of even a single alpaca would be very significant. It is thus clear that the use of Great Pyrenees dogs to guard livestock is an effective and very longstanding agricultural practice.”

Here there was a significant question as to whether the dogs were barking that much in any case.  They appear to have only barked at night, and were often no louder than many ambient noises in the area.  Predators are for the most part nocturnal.

These cases at least acknowledge that livestock guarding dogs often operate without handlers who can stop them from barking.  Barking is a deterrent to predators.  If dogs can bark despite the nuisance aspect of the sound, it must be accepted that they will do so when their protectiveness leads to bites.  Should owners train the dogs to attack predators but not attack people, even if people are threatening? Sheep rustlers, more common with the rise in lamb prices, is a real threat.  Is teaching guarding dogs not to react to people even practical?  Some degree of acclimation to humans may help, but too much familiarity is likely to reduce the protective aggressiveness required for the dogs to drive away prey animals. 

Interbreeding of Wolf and Shepherd Dog Populations in Georgia

Republic of Georgia (U.S. State Department)
Some analysis of the history of livestock guarding may also add perspective.  Four researchers from Ilia State University in Georgia (Kopaliani et al., 2014) have found that hybridization between wolves and livestock guarding dogs is a “common event in the areas where large livestock guarding dogs are held in a traditional way,” though such hybridization has probably declined since “humans started to more tightly control contacts of purebred dogs.”  Livestock guarding dogs are, according to these authors, widespread throughout Anatolia, the Caucasus, and mountainous parts of Iran, Iraq, and Turkmenistan.  The researchers found gene flow from dog to wolf populations and from wolves to shepherd dogs.  They note: 

“The majority of publications describing dog–wolf hybridization suggest that mating between male dogs and female wolves is more common than the other way around …. However, genetic analyses suggest that hybridization between male wolves and female dogs also occurs in nature …. The presence of dog maternal lineages in wolf populations … or sharing haplotypes between wolves and dogs can only be the result of such hybridization pattern: Female dogs can produce offspring both in the wild and in domestic conditions, whereas female wolves can breed mostly in the wild. Our study supports this point of view. Hybridization between male wolves and female dogs might happen both occasionally and deliberately: In mountain parts of Georgia, dogs are occasionally paired with captured wolves, which allegedly ‘improves the breed.’ In such deliberate hybrid occasions, both male and female wolves can participate. The latter case is the most likely explanation of shared haplotypes between dog and wolf.”

Thus, the interbreeding of wolves and dogs is not only accidental, but sometimes deliberate.  There was evidence of “recent wolf ancestry in more than 10% of shepherd dogs and recent dog ancestry in more than 13% of wolves.  Moreover, 2% of the studied wolves and 3% of dogs were, with a high probability, first-generation hybrids.”  The researchers note that studies of hybridization in Europe generally do not find such high levels of hybridization, probably due to the fact that in most of Europe dogs are more highly controlled.  They note:

“Large livestock guarding dogs, such as Great Pyranees, are not commonly used any more in a way that they can easily interbreed with wolves, but nobody can say to what extent they interbred with wolves in the past….   We hypothesize that the situation was much more flexible in earlier times, when most of the dogs used by common people were not subjected to intensive selection, similar to what is now the situation with livestock guarding dogs in the Caucasus and most likely the rest of West Asia. This means that interbreeding with gray wolves was an important part of the dog maintenance, and the situation was much more complicated than the simple pattern including Neolithic domestication with the further expansion of dogs descending from these early domesticated wolves.” 

The latter conclusion puts this research into the debates regarding domestication models, previously discussed here in other blogs. 

Interbreeding Wolves and Sheep Dogs in Antiquity  

Cylinder Seal Showing Sheep Dog with Double Collar, Lion, and Shepherd (Henri Frankfort, 1955)
The problem of aggressive livestock guarding dogs may be ancient.  Guarding sheep  is one of the earliest known functions of dogs, and is mentioned in Homer and the Book of Job.  Even in antiquity, there was concern that sheep guard dogs could become too vicious from intentional interbreeding with wolves.  The Mishnah, a Jewish text composed in the late second to the early third century, specifies in Kilaim 1.6 that the wolf and the dog, and the wild dog and the jackal [הזאב והכלב כלב הכופרי והשועל], are of diverse kinds, though they are similar. Merlen (1971) took this to mean that neighboring peoples were breeding dogs with wolves, presumably to produce more fearsome guard dogs, a practice the rabbis wanted to discourage among Jews, arguing that it was a violation of Leviticus 19:19 (“You shall not allow two different kinds of beasts to mate together.”). Kilaim 8.6 states that the dog is a kind of wild animal, though Rabbi Meir says it is a kind of cattle, i.e. domesticated. So breeding a dog with a wolf would be mixing a wild animal with a domesticated one. 

Perhaps these mixed dog-wolves were dangerous to travelers at a time when people walked the roads.  A Middle Eastern cylinder seal from before 1,000 BC shows a sheep dog attacking a lion so that the shepherd can put a spear in the lion’s back from the other side.   

Interbreeding of dogs with wild canids was also known in pre-Columbian Mexico.  Valadez et al. (2002) discusses the "tradition of breeding female dogs and male coyotes, because the hybrids were considered as resistant, loyal but temperamental, and good guardians."  Evidence from excavations at Teotihuacan, north of Mexico City, indicates that dog-wolf hybrids may have been bred specifically for ritual purposes, kept isolated from other animals and fed a vegetarian diet.  "We propose that the hybrids were created by men, with the purpose of having individuals that could display the sacred attributes of the father (wolf), inside a body that could be manipulated by men. These animals could be taken care until the sacrificial ceremony ... in which they were cut apart and separated in pieces according to the symbolic value.  The most important body parts were the heads and faces, which were used for priest and warrior attires." 


There is no doubt that livestock guarding dogs are an essential part of the sheep industry.  A study of livestock guarding dogs in the French Alps concluded that up to 95% of kills by predators can be prevented by combining the use of livestock guarding dogs with a practice of penning sheep at night.  Espuno et al. (2010).

The prior blog on this matter suggested that the sponsors of the bike race should perhaps have verified that there were no livestock guarding dogs in the area.  The owners of the sheep should perhaps be careful to keep the sheep away from roads where bicyclists and hikers may be encountered.  The forest service should perhaps serve as an information conduit between groups that will overlap on their use of forest lands.  These, however, are planning points, not resolutions of the legal issues involved here. 

Perhaps general concepts of negligence should be applied.  Was it foreseeable that the sheep were grazing in an area where humans were likely to be encountered?  Was it foreseeable that the dogs were so estranged from humans that they would see any animals, including humans, as threats to the herds they were protecting?  Should the sponsors of the bike race be liable for not investigating the dangers that might lie on the path of the race?  Should the dog owners have been more aware of the aggressive tendencies of their dogs?  Should they have considered that deploying the two dogs together might have increased their aggressiveness?  Had the dogs, 9 and 11 years old, become more aggressive with age?

Again, such questions may resolve a dispute without really interpreting the statute or finding the legislative intent behind its creation.  In the end, we are of the opinion that the issue of control has to be based on the context of the training and function of livestock guarding, which means that if the dogs were trained to drive away and if necessary engage with prey animals, they must be allowed to do their work if the statute permits their existence at all.  Applying concepts of control that have been used with other types of trained and working dogs will not produce a practical result for the livestock industry.  The dogs were in the owner’s control if they acted within the parameters of their training and purpose. 

This blog was written by John Ensminger and L.E. Papet


Blum, Karen A. (2001).  Saying “Neigh” to North Carolina’s Equine Liability Act.  North Carolina Central Law Journal, 24, 156 (discussing use of llamas to protect sheep: “llamas are extremely effective guard animals. One llama can protect up to 2000 sheep. Although llamas cost more than dogs, they cost less to maintain because llamas eat the same food and receive the same vaccines as sheep.”).

Espuno, Nathalie, Lequette, Benoit, Poulle, Marie-Lazarine, Migot, Pierre, and Lebreton, Jean-Dominique (2010).  Heterogeneous Response to Preventive Sheep Husbandry During Wolf Recolonization of the French Alps.  Wildlife Society Bulletin, 32(4), 1195-1208. 

Fisher, Kristina Gray (2008). Reclaiming Querencia: The Quest for Culturally Appropriate, Environmentally Sustainable Economic Development in Northern New Mexico.  Natural Resources Journal, 48, 479 (New Mexico farmers bought livestock guard dogs from the New England Farm Center in Amherst, Mass., which reduced sheep losses from 45% to 12% in one summer).  

Frankfort, H. (1955). Stratified Cylinder Seals from the Diyala Region.  The University of Chicago Oriental Institute Publications, LXXII. University of Chicago Press, Chicago.  (The interpretation of the dog as assisting the shepherd is mine, not Frankfort's.  Frankfort describes that lion as being "driven off in the nick of time with a spear, while his dog has growlingly retired before the formidable robber."  I doubt that either humor, or cowardice, is being depicted here.  Rather, I believe the seal shows their cooperation.)

Kopaliani, Natia, Shakarashvili, Maia, Gurielidze, Zurab, Qurkhuli, Tamar, and Tarkhnishvili, David (2014). Gene Flow Between Wolf and Shepherd Dog Populations in Georgia (Caucasus). Journal of Heredity, 105(3), 345-353. DOI: 10.1093/jhered/esu014.

Lorenz, Jay, and Coppinger, Lorna (2002). Raising and Training a Livestock-guarding Dog, Oregon State University Extension Service.

Merlen, R.H.A. (1971). De Canibus. J.A. Allen & Co. Ltd., London, 38.

Randi, Ettore, Hulva, Pavel, Fabbri, Elena, Galaverni, Marco, Galov, Ana, Kusak, Josip, Bigi, Daniele, Bolfikova, Barbora Cerna, Smetanova, Milena, and Caniglia, Romolo (2014).  Multilocus Detection of Wolf x Dog Hybridization in Italy, and Guidelines for Marker Selection.  PLOS One, 9(1), e86409 (looking at dogs and wolves in Italy, the Balkans, and the Carpathian Mountains, and finding hybridization “mostly attributable to village dogs and not strictly patrilineal.”  Not finding many F1 or F2 hybrids, the team says that their results suggest “that hybridization events already occurred in Italy some generations in the past.”).

Strack, Hermann L., and Billerbeck, Paul (1922) Das Evangelium Nach Matthäus.  Munich: C.H. Beck’sche Verlagsbuchhandlung (noting, at 722, the use of the term “village dog”, “Dorfhunde” in German).     

Valadez, Raul, Rodriguez, Bernardo, Manzanilla, Linda, and Tejeda, Samuel (2002). Dog-wolf Hybrid Biotype Reconstruction from the Archaeological City of Teotihuacan in Prehispanic Central Mexico.  9th ICAZ Conference, Durham 2002.  Dog and People in Social, Working, Economic or Symbolic Interaction (Snyder, L.M., and Moore, E.A., eds).


  1. hi-
    i just wanted to point out that humans can be malevolents as well, especially in this day and age (stealing sheep, killing or maiming sheep--for reasons of sport, religious rite, food, vandalism, etc) and thus, the guarding of the sheep against humans should be a necessary part of the sheep guarding dog's job and training. while a bicycle rider on a main road (was it on a main road or was it off-road?) during a race isn't necessarily an ill-intentioned human (although they could be), a guard dog won't necessarily know that--especially if the rider is coming too close to the sheep.

    if the human was nowhere near the sheep and a sheep guarding dog attack occurs, there may be less of a justification for guarding the sheep and more justification for an unreasonable attack.

    1. Good comment @tuffy! Since the Guarding dog's are carrying a lethal weapon at all times and using at will if they suspect a threat, it would seem reasonable for human's to also carry a lethal weapon for use if they suspect a threat. I think it's time to advocate for the safety of our brother's and sister's, not someone's bottom line.