Human-Wildlife Interactions 5{^):^06-^^^, Spring 2011

Good fences make good neighbors: imple-mentation of electric fencing for estab-lishing effective livestock-protection dogsT H O M A S M . G E H R I N G , Department of Biology, Central Michigan University, Mount Pleasant, Ml

48859, USA gehri1tm@cmicii,eduK U R T C . V E R C A U T E R E N , USDA/APHIS/Wildlife Services' National Wildlife Research Center,

Fort Collins, CO 80521, USAA N N A C . C E L L A R , Department of Biology, Central Michigan University, Mount Pleasant, Ml

48859, USA

Abstract: To be effective, livestock protection dogs (LPDs) must be carefully integrated withthe livestock they will be protecting. Others have developed guidelines to assist producersin this training and assimilation process. In many areas fencing is necessary; however,guidelines for containing LPDs and their livestock behind electric fencing are lacking. Wepresent results from larger projects involving LPDs where we encountered issues withfencing and preventing LPDs from roaming from their owner's property. We found thatranging and escaping from pastures was exhibited by LPDs that were not properly introducedand trained to electric fencing at an early age. LPDs that were trained strategically torespect electric fencing were effective guardians and did not leave livestock pastures.

Key Words: Canis lupus, depredation, fencing, human-wildlife conflicts, livestock protectiondog, Michigan, wildlife damage management, wolf

LIVESTOCK PROTECTION DOGS (LPDS) havebeen used in Europe for centuries, but onlysince the 1970s in the United States (Coppingerand Coppinger 2001). LPDs are an eftectivemethod for reducing livestock losses fromcoyote {Canis latrans) prédation (Green et al.1984, Coppinger et al. 1988, Andelt 1992, Andeltand Hopper 2000) and may have application foradditional human-wildlife conflict situations,such as prédation by wolves {Canis lupus) andmitigation of disease transmission from wildlifeto livestock (Gehring et al. 2010). Explicitguidelines for implementing LPDs on farmswhere it is necessary to contain livestock andLPDs behind fences are not readily available.General guidelines outline the basic principlesof integrating an LPD into a working livestockoperation (Lorenz and Coppinger 1988, Greenand Woodmft 1999, Dawydiak and Sims2004). Traits and behaviors (trustworthiness,attentiveness, and protectiveness) to look forin eftective LPDs and the process of socializingdogs to livestock (mainly sheep) are established(Coppinger et al. 1983, 1987; Lorenz andCoppinger 1988). However, specific instmctionson several aspects integral to the success ofLPDs are missing. The exact process for usingelectric fencing for integrating LPDs intopasture systems is 1 such missing component.

The dearth of information on how to use

electric fencing to manage LPDs could be dueto several reasons, including that LPDs are arelatively new tool for U.S. livestock producersand have not received widespread attention.LPDs also have most commonly been used inthe western United States where open-rangegrazing exists and containment is less integralthan it is in other parts of the United States.As gray wolf populations continue to expand,small- and medium-sized farms prevalent inthe Great Lakes region of the United States havea need to protect their livestock from prédationusing nonlethal methods, such as LPDs, whichare more preferred by society (Gehring et al.2010). To implement LPDs eftectively, producersmust be able to integrate them into their normalfarming practices, which usually involve theuse of fencing for containing livestock.

To be eftective, LPDs must defend livestockfrom predators, and, to do so, they must staywith livestock. Current LPD guidelines stress theimportance of bonding dogs with the livestockthey are to protect (Lorenz and Coppinger1988). Thus, most guidelines imply that roam-ing behavior by LPDs can be avoided by meansof proper socialization with livestock. Althoughstrong socialization is paramount for success,socialization alone may not prevent dogs fromroaming. An LPD is not protecting livestock if itis roaming away from them. Roaming also could

Good fences • Gehring et al. 107

lead to increased mortality of LPDs from suchcauses as vehicular accidents, shooting due totrespass, unintentional poisoning, or prédationby wolves or other predators. Roaming mightbe more common in small or suburban-fringefarms where properties are smaller andthere are more human-related distracfions oractivities to entice LPDs to leave their livestockor pasture (Green and Woodmft 1990, Gehringet al. 2010).

Using woven-wire fencing, chaining dogs,and using chain drags on dogs have beensuggested as techniques to reduce roaming(Dawydiak and Sims 2004). VerCauteren et al.(2008) used invisible-fence systems to containLPDs in 1.2 ha research pastures. Neutering orspaying LPDs may also limit roaming (Greenand Woodruft 1990). Culling individual LPDsthat roam may be necessary if a persistentproblem exists (Lorenz and Coppinger 1988,Green and Woodruft 1990). Although manyfarms incorporate some type of fencing intotheir normal husbandry practices, fencingrequirements for LPDs can be distinct fromfencing needs of livestock. No currentguidelines, however, specify fencing needs ofLPDs. As part of a larger study of LPDs over a5-year period, we found that eftective fencingand training was a crucial link for successfullyincorporating LPDs into working farms andpreventing roaming. We provide observationsand guidelines on fencing requirements andtraining needed to reliably contain LPDs andkeep them within pastures and with theirlivestock.

Training and integratinglivestock protection dogs

During 2005, we placed a male (neutered)and a female (spayed) Great Pyrenees pup (7 to8 weeks old) on each participating cattle farmin the Upper Peninsula of Michigan as partof a larger study examining eftectiveness ofLPDs to reduce livestock losses from predatorsand bovine tuberculosis. We purchased thepups from a reputable breeder of workingLPDs to ensure consistent behavioral traits indogs, and all pups were full- or half-siblings.Farms contained 19 to 50 head of cattle on10- to 40-ha pastures. We provided producerswith guidelines, and with our assistance,the providers were responsible for care and

Figure 1. Great Pyrenees pups at the time of deliv-ery to study farms for training as livestock protectiondogs.

training of pups. Pups were raised in 2 x 4-mpens (LPD pen) within an 8 x 8-m livestock penwith 2 <l-week-old calves. Under producersupervision, any negative behavior such aspulling tails or playing too rough was corrected.We provided food and water in the LPD penwhere only the dogs had access. Human contactwas minimized to ensure LPDs were bonded tocaftle and not humans. To avoid injury to pupsor any negative interaction between pups andadult cattle, pups remained with calves andhad contact with adult cattle at 4 to 7 monthsof age under supervision by producers only.Pups were allowed to be loose inside caftlebams or for short periods of time in pastureswhere they would be guarding caftle. Theywere kept on leashes when away from caftle ortheir pastures.

We began integrating the pups into adultcaftle herds when they were 7 months old. Atthis time, we housed pups and calves duringdaylight hours within outdoor pens ( 5 x 5m) inside pastures. Pups were walked dailyon leashes around the inside of pastures tofamiliarize them with pastures and to establishthe fence as a boundary (Figure 2). Pups wereencouraged to interact with adult caftle whileexploring pastures. We maintained this schedulefor 10 to 14 days, after which pups and calveswere housed in outdoor pens continuously for10 to 14 days. The dogs were then released intopastures. This slow-release program allowedpups to become accustomed to living in a newarea while furthering bonding between themand adult caftle.

Before pups were released into pastures, we

108 Human-Wildlife Interactions 5(1)

Figure 2. Young livestock protection dogs onleashes are walked around the fence perimeter.

added strands of 12-gauge electric fence wire foexisfing fence af each farm fo help confain pupsand prevenf fhem from roaming. We placed 1electrified sfrand of wire approximafely 0.25m above fhe ground at each farm. Addifionalsfrands were added as needed fo exisfingfences af each farm; a sfrand was added wheregaps between fence sfrands were >0.33 m. Wemonifored LPDs and fencing regularly andmaintained electric fencing current at 7,000volts. If an LPD exited a pasture, we identifiedany escape points and added strands ofelecfrified wire as needed. Also, we attemptedto correct escape behavior by pushing LPDsinto the electric fence fo ensure fhaf fhey had anegafive associafion wifh if. If escape behaviorcould not be prevented with electrified fencing,we implemented an invisible fencing sysfem(PefSafe Sfubborn Dog Sysfem, Radio SysfemsCorp. Auburn, Ind.). We placed invisible fencewire outside livestock fencing (eifher lying onfhe ground or hung on fence posts), aroundthe entire perimeter. We buried invisible fencewires at gates so that it was convenient forproducers and vehicles would not damage thewire. We fitted dogs with an electronic shockcollar that transmitted a corrective shock whenLPDs were within 0.30 m of the fence. After anintroducfion to the invisible fence, dogs wereagain released into pastures following the samesteps for electrified wire. The existing fencesserved as visible boundaries, so LPDs requiredlittle training once outfitted with electroniccollars. Our research was approved by theInsfifufional Animal Care and Use Commifteeat Central Michigan University (IACUC #13-04).

ObservationsSocialization

LPDs on all farms were well-socialized andonded wifh caffle. As pups, fhey would sleep

: inmediafely nexf fo fheir calves (i.e., trustwor-thy behavior). When pups were released fromfheir pens fo exercise or play in fhe bam, fheywould regularly be found greeting or inferacf ingwifh calves fhaf had once shared fheir pens (i.e.,affenfive behavior). Once adulf cattle becameaccustomed fo the presence of LPDs, fhey tendedto greet LPDs or freely graze without beingdisturbed by fhem. Once released into pastureswifh adulf caffle, LPDs would greef caffle in asubmissive fashion, licking around fhe caffle'smoufh. Caffle became equally inferesfed inLPDs and would suck on fheir ears and collarswifhouf fhe LPDs exhibifing any aggressiveor play behavior. Even in pasfures, LPDs andcaffle would be found resfing and sleepingfogefher. LPDs also exhibifed profecfivebehaviors by barking af and chasing deer,predafors, and ofher wildlife from pasfures.

Fencing and roaming behaviorThe bond between LPDs and fheir livesfock,

along wifh fhe electric fencing, was sufficienffo keep dogs within pastures on 3 of 6 sfudyfarms. Although well-bonded to cattle, LPDson 3 farms habifually escaped from pasfures.After fhese LPDs inifially were released infopasfures, they remained with cattle. After <2weeks, these LPDs began testing the fence andescaping. When escape poinfs in the fence werediscovered, we modified the fence by addingwires fo specific secfions. Alfhough fencemodificafions prevenfed LPDs from escapingat that specific point again, LPDs would findanother exit from pasfures. Adding additionallines of fencing f o escapee farms prevenfed LPDsfrom roaming for <2 days. LPDs would findnew escape points in the fence and continue toroam. On 2 farms, LPDs would roam buf cameback fo fhe farm house or barn yard wifhin 1 fo2 days. While fhis was commonly the case, onthe third problem farm, LPDs went several kmfrom their farm and were found by neighborson several insfances.

We abandoned use of elecfric fencing atthese 3 farms and implemenfed invisible fencefechnology. Immediafely, fhe invisible fencing

Good fences • Gehring et al. 109

system was successful in correcting roamingbehavior of LPDs. Within the first year of useof invisible fence on 1 farm, a break in the linecaused the system to fail for 2 days. Duringthis time, 1 LPD escaped after neighborsbegan shooting at birds near the pasture. Thisdog escaped through the fence and joined theneighbors. No LPDs escaped over the remaining3 years of study after the invisible fence wirewas fixed.

After extensive interviews with all producersin our study, we learned additional informationabout variability in producer training of LPDpups and deviation from our guidelines. Onall 3 farms where LPDs escaped from pastures,producers allowed dogs as pups to move inand out of pastures under an electric fencenear their barns (i.e., before we added linesof electric fencing for release of LPDs intopastures). Because fencing at all 6 farms wasdesigned for cattle, none had an electric wire<0.3 m from the ground. Pups could easily crawlunder and even between fence lines withoutconsequence. Although all producers knewLPD pups should not be allowed to breachthe fence when in pastures, 3 producers werecareless regarding this guideline. They allowedthe pups to roam unattended, often for hours ata time. We observed 1 producer calling his pupsto come to him when he was outside the fence.Initially, the LPDs stopped when they reachedthe fence line. However, after being called andcoaxed by the producer, the pups breachedthe fence to get to their owner. Allowing andencouraging this type of behavior led to LPDsnot respecting the fence as a boundary. Despiteour immediate actions to correct producerswhen they did not follow our protocol, it islikely that some producers continued to notfollow our guidelines.

On 3 farms where LPDs did not escape frompastures, we found that producers had alwayssupervised LPDs until they were integratedinto pastures with adult cattle. LPDs wereallowed to play inside bams and in smallpens. One farm allowed their dogs to be looseoutside with supervision in the barnyard. Noneof these farms, however, allowed their dogs tocross a fence or go into pastures unless enteringthrough a gate on a leash. Because LPDs werenot exposed to the fence until extra lines wereadded, their first experience with fencing was

negative. On these farms, we did not observedogs testing fences after initial exposure to it.Further evidence that these LPDs recognizedthe fence as a boundary and did not test itoccurred at multiple times throughout thestudy. For example, two of the farms had theirfence lines broken and one had a gate down for>3 days, but LPDs never left their pastures.

Management recommendationsFencing that prevents LPDs from leaving

livestock that they are protecting is a criticalcomponent in establishing eftective LPDs. Wefound that proper socialization alone was notenough to ensure that LPDs remained with theirlivestock. However, LPDs that begin roamingmay still become excellent guardians if theycan be properly re-conditioned and contained.We found that invisible-fence was eftective incontaining LPDs that previously had roamed.Both electric fencing and invisible fencinghave advantages and disadvantages associatedwith them. The type of fence used depends onexisting infrastructure and the needs of eachindividual producer. Both fence types can belabor intensive, requiring proper maintenance(e.g., monitoring downed wires, mowing orherbicide treatment for weed growth, andaddition of extra wire). However, electricfencing maintained for livestock (particularlysheep) may require only slight modificationsfor preventing LPDs from escaping. We foundthat the lowest electric wire (preferably <0.25 mfrom the ground) was the most critical. Invisible-fence technology is limited to approximately40 ha that can be fenced. Invisible fencing alsorequires regular monitoring to fix line breaksthat may occur during freeze and thaw periods,replacement of batteries in shock collars,and monitoring of LPDs to be sure that theirnecks are not being abraded from shock collarprobes. However, invisible fencing or 1.2-m-tallelectrified woven wire may be the only viableoption if producers are not completely investedinto training their LPDs. We found that a 610-msection of electric or invisible fence in our studycost $0.11/m and $0.63/m, respectively. Ourfarms required <3 additional strands of electricfencing at an estimated cost of $0.33/m.

Regardless of the type of fence used, trainingLPDs as pups to respect electric fences iscritical to establishing the fence as a physical

110 Human-Wildlife Interactions 5(1)

and psychological boundary and reducing thelikelihood of roaming behavior. We suggest thatthis initial training can be even more criticalthan the choice of fencing itself. Establishinga fence as a boundary by walking LPDs on aleash around a pasture's perimeter and, mostimportantly, not allowing them to pass througha fence is a crucial component of LPD training.Making an LPD's first experience with the fencea negative one is important in keeping themin pastures. During a LPD's first exposure toelectric fencing, we recommend leashed LPDsbe allowed to investigate fencing and receive ashock, followed by confinued walking of fenceand shocking. To reinforce aversion to fencing,producers should walk LPDs daily aroundpastures for several weeks, with correctiveshocking, before releasing LPDs into pastures.

Fencing and training guidelines1. Assess existing fencing on livestock

farm. Add additional electric wire, ifneeded, and ensure bottom wire is <0.25m from the ground. Altematively, usean invisible-fence system, woven wireor woven-electric net fencing. Electricfencing or invisible fencing is mostdesirable to prevent LPDs from diggingunder fences and escaping.

2. Socialize and bond LPDs with livestockstarting at 7 to 8 weeks of age. Ensurethat LPD pups are not allowed to breachelectric fence and that first experiencewith fencing is negative.

3. Integrate LPDs into pastures atapproximately 7 months of age. Ensurethat this integration is a formal processof walking dogs on leashes around thepasture perimeter and forcing dogs intoelectric or invisible fencing. When LPDsdemonstrate great resistance to beingforced near fencing, they are conditionedproperly.

4. If roaming problems are exhibited byLPDs, producers can use invisible-fence,in lieu of immediately considering cullingto correct this behavior.

5. Regardless of fencing considerations,producers must be genuinely interestedin using LPDs and be full participantsin training and maintaining dogs. If aproducer lacks these attributes, LPDsshould not be used.

AcknowledgmentsOur work was funded by Central Michigan

University (Research Excellence Fund Award),Central Michigan University College ofGraduate Stiidies, USDA/APHISAVildlifeServices' National Wildlife Research Center,USDA-Sustainable Agriculture Research andEducation Program, Defenders of Wildlife,Animal Welfare Institute (Christine StevensWildlife Award), CITGO Petroleum Inc., andNational Geographic Society-ConservationTrust. We are grateful to livestock producersand students who participated in this study. Weespecially thank M. Provost for valuable fieldassistance throughout the project.

Literature citedAndelt, W. F. 1992. Effectiveness of livestock-

guarding dogs for reducing prédation on do-mestic sheep. Wildlife Society Bulletin 20:55-62.

Andelt, W. F., and S. N. Hopper 2000. Livestockguard dogs reduce prédation on domesticsheep in Colorado. Journal of Range Manage-ment 53:259-267.

Coppinger, R., and L. Coppinger. 2001. Dogs: anew understanding of canine origin, behavior,and evolution. University of Chicago Press,Chicago, Illinois, USA.

Coppinger, R., L. Coppinger, G. Langeloh, L. Get-tler, and J. Lornez. 1988. A decade of use oflivestock guarding dogs. Proceedings of theVertebrate Pest Conference 13:209-214.

Coppinger, R., J. Lorenz, and L. Coppinger. 1987.New uses of livestock guarding dogs to reduceagriculture/wildlife conflicts. Proceedings of theEastern Wildlife Damage Control Conference3:253-259.

Coppinger, R., J. Lorenz, J. Glendinning, and P.Pinardi. 1983. Attentiveness of guarding dogsfor reducing prédation on domestic sheep.Journal of Range Management 36:275-279.

Dawydiak, O., and D. E. Sims. 2004. Livestockprotection dogs: selection, care, and training.Alpine Blue Ribbon Books, Loveland, Colo-rado, USA.

Gehring, T M., K. C. VerCauteren, and J-M. Lan-dry. 2010. Livestock protection dogs in the 21stcentury: is an ancient tool relevant to modernconservation challenges? BioScience 60:299-308.

Green, J. S., and R. A. Woodruff. 1990. ADC

Good fences • Gehring et al. 111

guarding dog program update: a focus onmanaging dogs. Proceedings of the VertebratePest Conference 14:233-236.

Green, J. S., and R. A. Woodruff. 1999. Livestockguarding dogs: protecting sheep from preda-tors. U. S. Department of Agriculture, Agricul-tural Information Bulletin 588. Washington,D.C. USA.

Green, J. S., R. A. Woodruff, and T. T. Tueller.1984. Livestock-guarding dogs for predatorcontrol: costs, benefits, and practicality. Wild-life Society Bulletin 12:44-50.

Lorenz, J. R., and L. Coppinger. 1988. Raisingand training a livestock-guarding dog. OregonState University Extension Service, PublicationEC 1238. Corvallis, Oregon, USA.

VerCauteren, K. C, M. J. Lavelle, and G. E. Phil-lips. 2008. Livestock protection dogs for deter-ring deer from cattle and feed. Journal of Wild-life Management 72:1443-1448.

KURT C . VERCAUTEREN conductsresearch to create novel means of managing thediseases and damage of deer and elk to protectAmerican agriculture. He is a project leader for theUSDA/APHIS/Wildlife Services' National Wildlife Re-search Center. He received his B.S. degree from theUniversity of Wisconsin-Stevens Point and his M.S.and Ph.D. degrees from the University of Nebraska-Lincoln. Much of his efforts are focused at the inter-face between free-ranging cervids and livestock. Heis the chair-elect of the Wildlife Damage Manage-ment Working Group of The Wildlife Society.

THOMAS M . GEHRING is an associateprofessor of biology in the Department of Biology atCentral Michigan University. He obtained his B.S.and M.S. degrees in biology and wildlife from Uni-versity of Wisconsin-Stevens Point and his Ph.D.degree in wildlife ecology at Purdue University. Hisresearch interests are related to carnivore ecol-ogy and conservation, particularly as these relateto human-wildlife conflicts. He teaches courses inecology, wildlife biology and management, mammal-ogy, and landscape ecology.

A N N A C . C E L L A R obtained her B.S. degreein biology from Ohio University. She is currentlycompleting her M.S. degree in conservation biologyat Central Michigan University.

Copyright of Human-Wildlife Interactions is the property of Jack H. Berryman Institute and its content may not

be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.