} . ·.~

Relative effectiveness of guarding·-dogbreeds to deter predation on domestic

sheep in Coloraclo

"/!limn F. Andell

Abstract Predators kill significant numbers of domestic sheep in Colorado and the westernUnited States. Thus, I obtained data from 119 sheep producers who used 1 breed oflivestock guarding dog (either ,-\kbash, Great Pyrenees, or Komondor) in 1993 and 59producers who used multiple breeds (including Anatolian Shepherd) in 1995 to assessrelative effectiveness of breed of dogs to deter predation on domestic sheep inColorado. For producers using 1 breed of dog, estimates of ewe and Jamb mortalitiesto most predators in most types of sheep operations. value of sheep saved from pred­ators, and ratings of effecti\eness did not vary among breeds of dogs. Producers usingmultiple breeds of dogs rated .-\kbash as more effective than Great Pyrenees to deterpredation. /v\ore producers also rated Akbash as more effective than Komondors toc1eter predation by all predators combined and by coyotes I Callis latrallsL /\\ore pro­ducers rated Akbash as more aggressive. active. intelligent, Jnd ['aster than GreatPyrenees. ,V\ore producers also rated ,-\kbash as more aggressive. attentive. trustwor­thy, active, and faster than Komondors..-\natolians were rated as Taster than GreatPyrenees. Great Pyrenees were rated as less active thJn Komondors. ,\.\ost producersindicated that the most important attributes of dogs were high aggressiveness to pred­ators. great attentiveness to sheep, and great trustworthiness, whereas fewer produc­ers indicated that great acti\ itv level, great inte!ligence. and fast mobility \\ ere impor­tant attributes.

Key words Akbash dog, animal damage management, black bears. Canis latralls, co\ otes, Pumaconcolor, Great Pyrenees, Komondor, mountain lions, predation, L'rsus americJnus

Pred:trors killed -i1.8% of the domestic sheep and!Jmbs [h:lt died in the l:nited St:ltes during 1994(National Agricultur:ll Statistics Service 1995).Guarding dogs h:lve been used successfully toreduce predatory losses on individu:l1 r:mches(linh:lrt et a!. 1979: Green et a!. 1984; Coppinger eta!. 19t>S: Green and Woodmff 1988; Andelt 1992;Andel! ,lI1d Hopper. in press). However. little d:ltaare a":li!:lble on the relative effectiveness of variousbreeds ()f dogs to deter predation. Green :mdWoodruff (1988) conducteL! a nation:ll sun'ey ofsheep .lI1d gO:lt producers and reported [hat effec­tiveness of Great Pyrenees. Komondor. Akbasl1,An:I[()li:lil..\bremm:l, ;\nd hybrids to deter [Jreda-

tion did nor vary. However. they reponed th;Jt moreKomondors th:ln Great Pyrenees..\.kbash. andAnatolians bit people. and signifiontly fewer Gre:ltPyrenees than Komondors, Akbash. and .\.natoliansinjured li\'estock. Green and \Voodruff (1983)pbced guarding dogs with sheep producers :lndreponed that Great Pyrenees were signitlontlymore successful than Komondors and Akbash to

Jeter prc:d:ltion un rangelands anL! pastures. Green:lnd Woodruff (990) also placed guarding dogswith sheep producers and reported that Great[>yrelH:es ~'ere more effeClive than :\natolians and[hat a gre:lter proportion of Anarolians injured ;\ndkilled liveslOck than did Great P~'renees. Green ;\nd

""~ildlife ~o~iC'ly Bulletin 19~9, 27(3):~O6-714 ...~. ~ r,~~"';.'-.... •~ _.• .. ,. w ... :o.. _.'

Pecr refer .c-l

-\kbash dogs guarding sheep on open range in Colorado. Photo'J\ William F. Andelt.

\,\'oodruff (1989) reported that .~bash and GreJ.tPyrenees both deterred b1:lck beJ.r (Lrslls ameri­

canus) predation on sheep. Coppinger et J.1.\.1983a. b: 1988) reported \Iaremmas were signifi­clJ1tly more attenti"e and more tru nvorthy than_·\.natOlians. In this paper. I report on the relati\'ecr"fectiveness of.~bash, Great Pyrenees. Komondor.Jnd ..\.natolian breeds to deter predJ.tion on domes­tic sheep in Colorado.

lVlethodsDuring 1993. I acquired complete or nearly com­

?!ete telephone sur;eys of 182 producers using;h'esrock guarding dogs in Colorado benwen 198-::­Jnd 1993 (.-\.ndelt and Hopper. in press). I contactedColorado StJ.te Cni"ersiry Cooperative ExtensionJgents in most counties of Colorado, examined pastrecords. and asked all guarding·dog owners sur­\'eyed for names of Ih'estock producers who wereusing or had used dogs during the last 5 years. [:mempted to contact all individuals named. Irequested information on rype of operation (fencedpasture. open range. feedlot); number of ewes andlambs in :m operation; ewe and lamb mortalitiesfrom all causes, all predators combined, and specif­ically from coyotes. black beJ.rs, mountain lions(Puma conc%r), and domestic dogs: breed andnumber of guarding dogs used: ratings of the: effe:c­tiveness of dogs; :.lI1d estimated monet:lry '·:.I!ue ofsheep 'aved by dogs during 1993 or the: laSt !'e:trthey used dogs. I discarde:d surveys of prouucers\vho used dogs only in feedlots, producers whoused dogs for < 1 year, :1Od producers who u. edmultiple breeds of dogs. but .-urveyed prodL CL:r~

with multiple bn:eds in more tkt:lil in I')')"i.

Great Pvrenees guarding sheep in Colorado. P:-.',·', :';' .....~i'iamF. Andelt

~umbers of ewes and lambs [hat were guarced bydogs were compared among producer,> wit:' differ­ent breeds of dogs using analysis ,;i ":lnance(.-\..."O\~-\. GL.\-I Procedure. 5.\5 Institute. :r:c. 1SiBS)after log transforming herd sizes to e~w..:Z~ "ari­ances.

:'\umbers of e'\.-es :.Ind lambs killed :"j~' ?:-e'::-,it'Jrsw<:re reported only for periods whe:; j,,:?~ ?'e:-ewith sh<:ep. \\11en dogs were not wit:: ,r:e=? :'orth<: entire Year. the mort:.Ilities were: e:c:-:;'7'~ ')~.. - . ~

dh'iding by the percentage Of the Jr.:-.'~ ::<7: r.tG

that dogs were with sheep. HO\\.T\-er. c;-;, '= ::.:.C ::W-:omortaliry rates to pred:.Itors "\."ere dele:!;'': :'.-: ~ ~rrr

Jucers because gU:lrding dogs ""ere :1','.. :;:~:- ,::e~

during lambing. '\."hen mort:liit:e,> !':', ::. :: :'::'~:',:-,

likely Jre greatest. I comp:lreu mort:;..: :'..-,"--.;:::~ .,<yJmong producers who indicned :1 ,r:-:'.:::<. ;-:e0.­tOr ,\."as found in the are:l of their ope:::.:: '.::. : ~';::-.­

pared numbers of e\ve and bmb :TJ(.r:,_:.-~c, :') ::.llpredatOrs combined and indi\'jduai]v ~'. _-_~. '.r.J::".blJck beJ.rs, mountJin lions. ;lnd rjr,r:.t=.,.-.Y"; -y,,,?,:

among respond<:nts with different hr~.-::-:''' ..... -r,;::-ing the slop<:s of regres ion iim:., .r:":<:.:-::r.::': .~.

regressing number of mortalities :.Igaln:,~. ::".:::::.i:-~ '-rewes or lambs maint:tined with rj(Jg.., : ·_·x..c·...tC.-..c:::ll1al~·ses with the GL.\[ procedure :.lfrJ::" ':',:""';"::-x-r-~

number of mortaliries by the recipr(,<.:, •. ':: ::.~~of ewes or lambs with dogs [() ~t.ahiliz,: ':'-.1"; :,::":--y:"~

v:triance in [he regression. \X'hen [he' ,'. >;::;_: ~--.J::-::;

indic;lted significant rre:trmen t d"fer.-., : _:~ 2­sided chi- 'quare tests with 1 Je;~n;e '.r ':o-:-;:".J·f::' -.-,

derermine which mortality rJ[e" c1iIf<::-::: :r::::''; *::':::

breeds. [considered:\ P<U.!J 1- .,igntr:( .:,_::.: :-::~

j-rf(;atmc:nt comp;lrisons to :TJ:.Iint.:J:r. .::. -::::.-;r--:;,..mc:nr·wisc error ralc_ or I!.I I").

I :t1~o Jcrermined cll:ll1gc:'> In ': ..~e ,r.': :~:""'l-: :~-:.

sizes and mortality rates from 1986 to 1993 amongbreeds for producers who did not have dogs in1986 but obtained them by 1993. I detennined per­centage changes in ewe and lamb herd sizes bysubtracting number of ewes or lambs owned in1986 from those in 1993, dividing by number ofewes or lambs owned in 1986, and multiplying by100. I compared these changes with A.'lOVA afterweighting percentage changes by herd sizes in1986. I determined changes in ewe and lamb mor­tality rates from 1986 to 1993 by subtracting theproportions of ewe or lamb mortalities in 1986from those in 1993 for individual producers. I com­pared these differences :lmong tre:ltments withA: OVA after weighting the differences in mortalityrates to :lll causes by mean number of ewes orlambs owned in 1986 and 1993 for each producer,where:ls mortality rates to pred:ltors were weightedby the mean number of ewes or lambs with dogs.

I used ANOVA to compare the reported monetaryvalue of sheep saved by dog breeds for producersusing 1 breed. I assigned numerical rankings to rat­ings of the effectiveness cf dog breeds (excel­lent=-!. good=3. f:lir=2, poor= 1, un:lccept:lble=O)and compared them with Al"lOVA using the GL\1procedure. I used the "protected" le:lst significantdifference test (~lilliken and Johnson 198-4:31-33,S:n'il!e 1990: 177) to determine which tre:ltmentpairs were different for average number of ewes:lnd lambs m:lintained by producers with dogs, per­centage changes in ewe and lamb herd sizes from1986 to 1993. differences in mort:lliry rJtes from1986 ro 1993. and producer ratings of effecth'enessamong breeds when the overJU F-test inJic:ned sig­niiic:lnt (P<O.05) treatment effects.

During 1995. I contacted, by telephone, :llmost allproducers who reported using multiple breeds ofdogs Juring 1993. I requested information on

Table 1. ~umber 01 ewes with dogs. percentage e\\e mortalities to all causes. and percentage ewe mortalities to predators in rela-tion to breed oi guarding dog used reported by Coiorado sheep producers using 1 breed 01 dog, 1993.

Characteristics 01 Akbash Great Pvrenees Komondor

sheep operations n x or 1)0 SE n x or Q~ SE n y or llri SE po

E\\es with dogs (x)

..33.-\bFenced pastures 9 160 53 33;",-\ 1~6 13 1"5A 60 0059Open ranges 1, 3.0;-3,-\ 565 15 ~.·F5A 933 a 0.139;\11 operationsC 31 ~.O~lC 386 ~ 3068 "- 13 1"5A 60 .;;0,001--~

..;11 causes 1%)

F~nced pastures 9 r •• 1.1 53 ".1A 0,5 13 ·l.OA ' . 0, I 590 ......., ,.\J

Open ranges 16 5.3A 0.9 15 6,5,-\ 1,0 a ,),631,\11 operations 30 ;,9,\ 0.5 ~ 5.;- .; 0,5 13 ",OA ' n 0.;-93-.0

,\11 predators 1.%1Fenced pastures 9 ~.: \ 0.6 53 0.5,-\ 0.3 13 1. ..A 0.9 0.05~

Open ranges 1, 1.1-\ 0,3 14 1,3A 0.3 a 0.118.-\11 operations 31 1,2-\ O,~ -.. 1,3A O.~ 13 1...A 1.1 0,396

Covotes ('Yo)Fenced pastures 9 2.18 0.6 53 0.3A 0.3 13 1."AB 0.9 0,032Open ranges 17 0,8A 0.3 1.. 1,5A 0.3 0 0.136All operations 31 0.9A 0,2 :-.. LOA 0.2 13 1...A 1.1 0.;-91

Black bears ('Yo)Fenced pastures 2 O.OA 0,2 18 O.lA 0.1 .. O,OA 0.3 0,905Open ranges 16 0.3A 0.1 10 0,2A 0,1 0 0.3;"5All operations 2~ 0.3.-\ 0.1 33 O,~.\ 0.1 .. O,OA 0.8 0.567

Mountain lions (%)Fenced pastures 3 O.OA 0.0 20 O.OA 0,0 7 O,OA 0,1 0,866Open ranges 13 O.lA 0.0 9 O.OA 0,0 0 0.327All operations ~O 0.1."- 0.0 32 O.OA 0.0 7 O,OA 0.1 0,2 .. 5

Dogs ('~"')

Fenced pastures 8 O.OA 0.0 .... O,IB 0,0 11 O,OAB 0.1 O.D21Open ranges 9 O.OA 0.1 8 0,38 0,1 a 0,0~6

1\11 operations 21 O,OA 0,0 5Cl 0.28 0,0 11 O,OAB 0.1 0,003

a Probability Ihat Ihe 3 means or re'~n,~sion slopes "" mortaliticsl in d row cio 11<)( tliiicr; numerJtor dl=2. denominalOr di=sum{Ji n's in J row minus 3.

b ,'"Ieal1s in a row iollowed by r1iiie:rt::nl letters diii<:r ,i:jI1,iicilI11Iy IP~O,lll ;"/col11p.lrisol1 (0 'Jbl.lil1 .111 e.\perimeI11-\\,ise J'~().OS),

c Includes iel1ceu pJ~lurc. open r.1n'-l<:. ienccJ p.l~lur<:-')pcn r.ll1!-l<:. iel1ced p.l';lur<:-'Jpcn r.ll1ge-iecdlot. JnJ it'l1ccdp.blure-ieedloIOperJlions.

. .,

Table 2. Number of lambs with dogs, percentage lamb mortalities to all causes, and percentage lamb mortalities to predators inrelation to breed of guarding dog reported by Colorado sheep producers using 1 breed of dog, 1993.

Characteristics of Akbash Great p\ renees Komondor

,neep operations n xor% SE n x or 0/0 SE n xor % SE pa

~Jmbs \\ ith dogs t.y!

6318bFenced pastures 9 187., -161.-\8 141 13 166A 64 0.042J_

Open ranges 16 3.39-1A 600 15 2.:-06.-\ 923 0 0.179-\11 operations C 30 2.290C 413 7-1 9668 233 13 166A 6-1 ~0.001

-\11 causes 1%)Fenced pastures 9 7.1A 1.7

., 5.0A 0.8 12 7.5A 2.7 0.399J_

Open ranges 14 8.9A 1.4 15 'i.7A 1.5 0 0.564-\11 operations 28 8.9A 0.8 'i3 6.5A 0.7 12 7.5A -1.1 0.082

-\11 predators (0\'1

Fenced pastures 8 2.2A 1.1 . .., 1.3A 0.5 12 -I.IA 1.8 0.267J_

Open ranges 15 -I.OA 0.8 1-1 3.9A 0.9 0 0.971-\11 operations 28 -I.OA 0.5 'i3 2.8.-\ 0.5 12 -I.2A 2.6 0.201

CO\OleS ,0\,)

Ft'nced pastures 8 2.2A 1.0 .,0.8A 0.5 12 -I.1A 1.6 0.081J_

Open ranges 15 3.7A 0.6 1-1 3.0A 0.7 0 0.-136-\11 operations 28 3.:-8 0.-1 73 2.1.-\ 0.-1 12 -I.2A8 2.2 0.016

31ack bears tOlo)

Fenced pastures 2 O.OA 0.6 18 0.3A 0.2 -I O.OA 0.7 0.872Open r.Jnges 15 0.3.-\ 0.1 10 0.1 -\ 0.1 0 0.137-\11 ooerations 21 0.3A 0.1 33 0.1.-\ 0.1 -I O.OA 0.8 0.166

\\ount~in lions ,°0

Fenced pastures 3 O.OA 0.6 20 0.1.-\ 0.2 7 O.OA 0.7 0.982O!Jen .Jnges 12 O.-IA 0.2 9 0.1.-\ 0.2 0 0.161-\11 0perations 19 O.-IA 0.1 32 0.1-\ 0.1 7 O.OA 1.0 0.091

::JOgS 00'

Fenced pastures 7 O.OA 0.3 -1-1 0.4.-\ 0.2 11 O.OA 0.6 0.-190Open ranges 9 0.0.-\ 0.0 8 0.0.-\ 0.0 0 1.000--\11 o:Jerations C 20 0.0.-\ 0.1 58 0.2.-\ 0.1 11 O.OA 0.5 0.374

.:! ?~cbability that ,he 3 means or regression slopes ,°'0 morralities' in a row do not dirier: numerator di=2 'except on open ranges.-:iT=, . cienominator di=sum of n's in a row minus 3.

::> \ le::ns in a ro\\ :ollowed b\ diiierent letters eifier significJnd\ P<O.Ol:- 'comparison 10 obtain In experiment-wise P<0.051.- ncudes ;'enced pasture. open range. fenced paswre-{)pen 'ange. :enced pasture-0pen ;ange-i·eedlot. and ienced pasture--

:eecloloperations.

breeds used: !.ltings of the relative effectiveness of\'3.rious breeds to deter predation by all predatorscombined and specifically coyotes. black bears.mountain lions. and domestic dogs; comparisons of\-arious beha\'iors among breeds; and the produc­er's rating of the importance of these beha\-iors todeter predation. I compared ratings of the relati\'eeffecti\'eness to deter predation and ratings of behaviorsof "arious breeds by producers using multiple breedswith the 'ign test using exact significance levelsfrom a cumulative binomial distribution table.r considered a P<O.05 'ignificant in all analyses.

ResultsDuring 1993, r used slln'eys from I 19 producers

who usetl only 1 breetl of dog (.-\.kbash, GreatPyn.:nees, or Komontlor). During 1995. I ootainctland used "urveys from 59 producers C22 in fencctl

pastures. 23 on open range. 11 in fenced pas­ture-open !.lnge operJtions. 2 on open !.lnge-feed­lor ope!.ltions. antl 1 in a fenced pasture-openrange-feedlot operJtion) that used multiple breeds(including Anatolian Shepherd).~llmber of ewes maintained with dogs did not

,·ary among breeds within fenced pastures and onopen ranges, but within all operations combined,producers maintained more ewes with.-\.kbash thanwith Great Pyrenees and Komondors and moreewes with Great Pyrenees than with Komondors(T:lble 1). Proportion of ewes lost to all causes. toall predatOrs combined, and individuaUy to b!:lckbears :lI1d mountain lions tlitl not vary amongbreetls. In fenced pastures, proportion of eweskilled b~' coyOtes was less for herds guarded byGreat Pyrenees compared to Akbash, whereas inkncetl pa 'wrc " on open ranges. or in all ope!.ltionscombined, rhe proportion of t:\ves killed by domes-

Table 3. 'umber of ewes with dogs, percentage ewe mortality irom all causes, percentage ewes killed by predators, and differ-ences in these parameters from 1986, before sheep producers had guarding dogs, to 1993, after producers obtained Akbash andGreat Pyrenees guarding dogs in Colorado.

Akbash Great Pyrenees

Characteristics of Percentage Percentagesheep operations n 1986 1993 change SE n 1986 1993 change SE pa

Ewes with dogs (x) 10 2.587 2,700 -1.-1 9.5 13 596 409 -31.-1 18.1 0.093,'vIortalities

All causes (%) 10 6.5 6.1 ·0.-1 1.0 13 5.8 4.5 -1.3 0.9 0.602All predators (%) 10 ., - 1.5 -1.0 0.5 13 1.2 0.3 -0.8 0.-1 0.854_.:J

Coyotes (%) 10 1.-1 1.0 -0.-1 0.1 13 1.0 0.1 -0.9 0.-1 0.298Black Bears (%) 9 0.2 0.4 0.1 0.1 8 0.1 0.1 0.1 0.1 0.;-53Mountain Lions (%) 9 0.6 0.1 -0.5 0.5 7 0.0 0.0 0.0 0.0 0.517

Domestic dogs (%) 5 0.5 0.1 -0.3 1.0 10 0.0 0.1 0.1 0.1 0.657

a Probability obtained by subtracting number oi e\\es or proportion oi ewes killed during 1986 irom those values ior 1993 iorindividual livestock producers and comparing the diiierences between ewes guarded by Akbash and Great Pyrenees.

tic dogs was less for herds gU:lrded by Akb:lsh thanby Gre:lt Pyrenees.

Within fenced pastures, producers maintainedmore !:lmbs with Akb:lsh th:ln with Komondors(Table 2). Within :lll operJtions combined. produc­ers m:lint:lined more lambs with _\kba h th:ln withGre:lt Pyrenees :lnd Komondors :lnd also had morelambs with Great Pyrenees than Komondors.Proportion of lambs lost to all causes: to JIl preda­tOrs combined: and indiYiduJUy to black bears.mountain lions. Jnd domestic dogs did not yary:lmong producers using Akb:lsh. Gre:lt Pyrenees.Jnd Komondors in fenced p:lstures. on openrJnges. or in JIl operJtions combined. but :l gre:lterproportion of bmbs WJS killed b~' coyores in herdsgU:lrded by Akbash than by Gre:lt Pyrenees in :llloperations combined.

PercentJge changes in numbers of e'l\'es Jndbmbs owned from 1986. before producers usedgU:lrding dogs, to 1993, Jfter producers used them.did not differ between herds gU:lrde<..l by .\kbash:lnd GreJt Pyrenees (Tables 3. of). Ewe and bmbmortalities to JIl CJuses. to JII predJtors combined.Jnd peciikJlly to coyotes. blJck beJrs. mountainlions, Jnd domestic dogs generJlly \vere less <..luring1993 than during 1986, but these differences inmortJliry rJtes did nor \'ary for sheep guarded b~'

_\kbash ;md GreJt Pyrenees. Producer estim:ltes ofthe monetary \-a]ue of sheep sayed did not \'J~'

Jmong breeds of gUJrding dogs in fenced p:lstures.on open ranges. or within :lll oper;ttions combined(Table 5).

Producers who use<..l only 1 breed of gUJrdingdog rJted .\kbash. Gre:lt Pyrenees, :lnd Komondors

Table -I. :"umber oi lambs with dogs. percentage lamb mortality irom all causes. percentage lambs killed by predators. and dii-ierences in these parameters irom 1986. ~eiore sheep producers had guarding dogs, to 1993, aiter producers obtained A"bash andGreat Pyrenees in Colorado.

."- bash Great Pyrenees

Characteristics of Percentage Percentagesheep operations n 1986 1993 change SE n 1986 1993 change SE pa

Lambs with dogs (x) 9 2./11 3,039 12.0 10.1 12 752 5-18 -16.;- 18.2 0.07-1'vIortalities

.-\11 causes (%) 'J 9.2 10.9 1.:- 1.3 12 5.9 8.0 1.1 0.9 0.3(,1

..\11 predators 1%) 8 :J., 6.-1 0.7 1.2 12 3.' 2.6 -1.3 0.8 0.219Coyotes 1%1 3 5.-1.... 5.7 0.2 1.1 12 3.8 2.-1 .1.-1 0.8 0.~fl5

Black Bears ('Yo) 7 0.1 0.6 0.5 0.2 7 0.0 0.2 0.2 n.2 1l.233;'v\ountain lions (0;\,) 7 0.1 0.9 0.3 O.J 6 0.0 0.0 0.0 0.0 1l.0;-6Domestic dogs (%) 5 (J.O 0.0 0.0 n.a (J o.n 0.0 0.0 0.0 10

.1 Probabllily obtained byublr.lLl,n·4 numb"r lli lamb~ (;( prl,plJrI,IJIl (,i !'lInbs krll"t! during 1986 irom Ihme v.1Iul:~ ior I'J') j iarindividu,11 livl"SIOck producers ,lnd LI;mparlng Ihe dlfi"'f:n<:l:cs !>el'sl;,:n I~mbs ~lI.Hderl by ,\kb:l,h ,1nd Creat Pyrenl"es.

Table 5. Producer estimates of the value (5) of sheep saved from predators by each guarding dog of 3 breeds in Colorado, 1993.

a Probability that the means for guarding dog breeds do not dirrer.

Characteristics of Akbash Great Pyrenees Komondor

,heep operations n x SE n ,y SE n x SE pa

Fenced pastures 7 2.138 947 41 l.3~6 332 6 417 242 0.339

Open range 13 3,763 1.307 12 5.1 ;8 2,162 0 0.574

... 11 operations 24 3,152 763 59 2,110 525 6 417 242 0.256

:IS similar in effectiveness to protect sheep from allpred:ltors combined and for specific predators(T:lbk 6). Both Akbash (P=O.O-iO) and Gre:ltPyrenees (P=O.O 13) were r.lted 3S more effective to

Jeter predation by cO~'otes compared to black.bears :lnd mountain lions in all oper:ltions com­bined. but the ratings of effectiveness for all 3breeds did not \':Iry among predators within fencedpastures and on open r:lnges.

:\lore producers who owned multiple breedsr:lted .-\kbash 3S more effective than Great Pyreneesro deter predation by all predarors combined :lndspecifically by coyotes. bl:Jck. belrs. mounrain lions.and domestic dogs (Table -:-)..\ gre3ter number ofproducers rated Akbash more effective than

Komondors to deter predation by all predatorscombined :lI1d by coyotes. Producers gener:lllyr:lted Anatoli3ns as more effective than GreatPyrenees; hO"'ever, small sample sizes did not allowdefiniti\'e conclusions. Akbash and Anatolians andGre3t Pyrenees. Komondors. and crosses betweenthese latter breeds generally were r:lted as similar ineffectiveness. but sm311 sample sizes do not allowdefinitive conclusions.

.\lore proJucers r:lted ,-\kbash as more :Iggressi\'e,3cti\'e, intelligenr. and fJster than Gre:lt Pyrenees(T3ble 8)..-\ greater number of producers also r:lted.-\kbash 3S more aggressi\·e. :menrive. trustworthy.3cti\·e. and f3ster than Komondors..-\narolians werer:lted 3S faster than Great Pyrenees. Great Pyrenees

Table ". Producers using 1 !Jreed or guarding dog Jnd ratings or criec:iveness or .. arious breeds :0 deter predators Irom preving::In domestic sheep in Colorado, 1993.

Characteristics or ,... 'bash Great ?\ renees Komondor

;heep operations n ya SE n ~ SE n y SE ph

Fenced pastures.-\11 predators 9 3.67 0.1/ -1 3.61 0.09 13 3.15 0.30 0.119J_

Covotes 9 3.56....c 0.18 53 3.64A 0.08 13 3.12,'" 0.32 0.071Black bears 2 3.50A 0.50 10 3."0A 0.2~ 1 3.00A 0.838,\.10untain lions 2 3.50A 0.50 9 3.56A 0.18 4 3.;-5A 0.25 0.805Domestic dogs 8 3.50A 0.38 40 3.30A 0.15 10 3.05A 0.34 0.612

Open rangesAll predators 17 3.38 0.21 15 3.50 0.24 0 0.715Coyotes 17 3A4A 0.19 15 3.53A 0.2 .. 0 0./62Black bears 16 2.69A 0.28 ;- 3.14A 0.34 0 0.361.Vlountain lions 8 2.63A 0.50 6 2.6;A 0.56 0 0.957Domestic dogs 8 3.00A 0.38 8 2.88A 0.48 0 0.841

,... 11 operationsAll predators 31 3.44 0.14 74 3.55 0.08 13 3.15 0.30 0.222Coymes 31 3.48B 0.12 75 3.60B 0.08 13 3.12A 0.32 0.102Black bears 22 2.82A 0.12 ~ 1 3.14A 0.19 1 3.00A 0.546.\.lountain lions 13 2.81A 0.36 16 1.13A 0.~6 4 3./5A 0.25 0,338Domestic dogs 20 3.35AB 0.22 54 3.17 B 0.14 10 3.05A 0.34 0./01

J ,Vleans were obtained by averaging producer ratings or the >:Ifectivcness r)r ~uarding dogs lexcellent=4. good=3. (;1ir=2,poor= 1. and unsati~lactory=OJ.

h Prob.lbdily IhJI :he mc,}1l'> lor guarding dog !>rer:ds do nrJt .!irl>:r..: .VlCJIlS ill .1 Lulur~n wllhin ienced pas;urcs. upr1/1 r.ll1~CS. ,1nd ,;11 r)Pcr.lI10ns lollOWl'd bv Ihe ~amc ICIIl'r do not dilfer.

Table 7. Number of livestock producers rating 1 livestock guarding-dog breed better or thesame as another breed to protect domestic sheep from predators in Colorado. 1995.

a The first number represents number of producers rating the first breed better than the sec­ond. the second number represents number producers rating the breeds equal, and the thirdnumber represents number of producers rating the second breed better than Ihe first breed,

b "=P<0.05, "=P<O,Ol, "·=P<O.OO1.

The proportion of ewe and !:lmb morulities tomost pred3.tors in most tvpes of sheep oper:Jtions,ch3.nges in mort3.lity rates before compared ro after

r.'

3

1'"

O'

E 2

o 3

-+ 6 2

3 0

17 4

7 0

5 0 2

6 9 3

Allpredators

, 0

o0"

1

2' 12 -; 6

3 -+ 2

o -+8 56 0

Domesticdogs

1 E 2

obt:lining guarding clogs,estimates of the v:I!ue ofsheep s:lved from preda­tors. and r:Jtings of effec­tiveness c1icl not differ:lmong Akb:lsh, GreatPyrenees, :lnd Komondorsas reported by producersusing 1 breed of dogs duroing 1993. These findingsconcur with :l n:ltion:llsur,ey tlut indiuted nodifference in effectiveness:lmong breeds (Green andWoodruff 1988). In con­tr:lst. producers in thisstudy who used multiplebreeds r:lted Akbash :lSmore effective than Gre:ltPyrenees :lnd Komondors

to deter pred:ltion. These findings differ fromGreen :lnd Woodruff (1983). who reported tlutGre:lt Pyrenees were more successful than .-\kbJsh.The differences between these studies I11Jy be rebt­ed to differences in str:Jins within breeds or per­lups selection for more effective .-\kb:lsh in recentye:lrs.

The l:J.ck of correspondence bem'een ratings ofeffectiveness of producers using ~ingle :lnd multi­ple breeds of dogs in this study indiutes that thedifferences noted by producers ~'ho used multiplebreeds :lre sm:lll or tlut comp:lrisons among pro­ducers who used only 1 breed are Jt too CO:lrse J

1

0"

ooo

E 2

o 0 0

o 0 0

Predator

1 0

10 3

2 1

2 01 1

Mountainlions

o

o

Blackbears

E 2

o

o

2a 2 1 1O···b 10 -+ l'

O' 2 0 10 4 1 01 1 3 0

4 6 2

o 3

Coyotes

E 2

2 217 5

7 04 3

4 13

Discussion

Guarding-dog breeds

Akbash. Anatolian

Akbash, Great Pyrenees

Akbash, Komondor

Anatolian. Great Pyrenees

Great Pyrenees. Komondor

Great Pyrenees. GreatPyrenees- Komondor cross

Komondor, GreatPyrenees- Komondor cross

were cued less :lcth'e than Komondors. .\lost pro­ducers felt tlut the most import:lnt :lttribures todeter pred:ltion were high :lggressiveness to pred:l­tors. high :lttentiveness to sheep. :lnd high trust­"':orthiness: some felt th:lt high JCti\'ity levels, intel­ligence. :lnd mobility were import:lnt attributes(T3.ble 9),

Table 8. :-.'umber of livestock producers rating 1 livestock guarding-dog breed's benavior above or the same as another breed whenprotecting domestic sheep Irom predators in Colorado. 1995. .

Bena\ior

,\I1ost .'v'ost .\I1ost Most .\I1ostaggressive attentive trustworthy active Fastest intelligent

Guarding-dog breeds 1 E 2 1 E 2 E 2 E 2 E 2 1 E 2

Akbash. Anatolian 3 0 3a 1 2 2 -+ 0 1 3 1 2 3 0 3 3 2

Akbash. Great Pyrenees 18 3 O···b 10 6 5 8 9 4 18 4 0'" 19 2 0'" 16 -+ 2"Akbash, Komondor 6 1 O' 7 0 O· 6 1 O· 7 0 O· 7 0 O· 5 2 0

Anatolian, Great Pyrenees 6 0 1 2 2 2 0 3 j 5 1 1 () 1 O' 4 2 I

Great Pyrenees, Komondor 3 7 8 8 5 5 12 2 4 2 4 11' j 5 8 5 IlGreat Pyrenees. Great

Pyrenees-Komondor cross 4 5 j I) 3 4 (, 2 2 4 (, 6 5 I) J

Komondor, GreatPyrenees-Komondor cross 3 0 2 ') 2 0 2 3 0 \)

J TI,e I'irst number represents number of producers ralin~ the lir;1 br(!"d betler than Ihe second. Ihe second number represenlSnumbl:r producers rating Ihe breeds ,"qu,ll, .jnd the third number repr,:;,'nlsl1l1l11ber III proullcers raling thr~ "er:ond hrr:r:d b"I1,'rthan Ihe lirst breed.

b ·=P<tJ.05, •• =P<O.O 1. • "=P<O,OO 1.

Table 9. Livestock producer ratings of the importance of various livestock guarding-dog behav­iors for reducing predation on domestic sheep in Colorado. 1995.

2

2

3

o1

o

otimportant

8

:>:>

12

less effective than GreatPyrenees.

Coppinger et a!. (1983a,b)indicated that attentive­ness, trustworthiness, andprotectiveness are themost important behav­iors of guarding dogs.Similarly, \1cGrew andBlakesley (1982) reportedthat aggressive and atten­tive dogs were more suc-cessful in protectingsheep. Livestock produc­ers in this study concurred.

Producers using multiple breeds of dogs rated.-\kbash :lS more Jggressive toward predators,:lccive, intelligent, :lnd f:J.ster than Great Pyrenees.Green Jntl Woodruff (1990) reported that GreatPyrenees (presumably compared to Akbash,,-\n:ltolians, and Kuvasz) were less likely to beJggressi\'e toward unfamiliar domestic dogs. Thus,the lower proportion of ewes killed by domesticdogs in herds guarded by Akbash compared toGreat Pyrenees for producers using 1 breed of dog:lnd the rJ.tings of greater effectiveness of Akbashcompared to Great Pyrenees by producers usingmultiple breeds probably were related to thegreater aggressiveness of .-\kbash. However. the,-\.kbash·s greater :lcti\'iry, intelligence. :lnd mobilitybehaviors :llso may have contributed. The moreaggressive. attentive. and trustworthy rJ.tings of,-\kbash compared to Komondors app:lrently attrib­uted to their rating of greater effecth-eness to deterpredation,

Green and Woodruff (1990) recommended using:lggressive breeds, such as All:ltolian..-\kbash andKomondor, where bears (Ursus spp.), mountainlions, and wolves are frequent predators. Producersusing multiple breeds of dogs concurred by ratingAkbash as more effective than Great Pyrenees:lgainst all predators. They also rJ.ted Akbash asmore effective than Komondors against all preda­tors combined :lnd coyotes.

Coppinger et :II. (l983a b; 1988) rated Anatoliansless in attenti\'eness and trustworthiness than.\-laremm:ls. Green :1Od Woodruff (1990) reportedthJt An:.ttoli:lns were rated less in performance thanGreat Pyrenees. with a gre:lter proportion of.\n:ltolians injuring or killing livestock. ProducersLlsing multiple hreeds of Jogs in my survey JI 0

tended [() rate' :\nato\ians less trustworthy th:.tn

Percentage

."'ediumimportance

Coymes .lrc the m.ljor predator oi dome,tic sh<:<:p n Cul,)racG.PhrAo by Willi,lm F.,\ndelt.

scale to determine relative effectiveness of differentbreeds. I surmise tInt rJ.tings of relative effecth'e­ness Jre most KCUrJ.tely portrJ.yed by producers'I\'ho used multiple breeds because the dogs gener­.Illy were used under similar conditions by individ­ual producers,

Producers using I breed of gUJrding dog and in.tll operJ.tions combined reported IJrger numbersof ewes and lambs guarded by .-\.kb:lSh than byGre:lt Pyrenees and Komondors. These differences.lpparently were related to larger herds of sheepgr.lzed on open range compared to those grazed infenced pastures :1Od to J preponderance of .-\.kb:lShused on open rJ.nge, whereas most Great PyreneesJnd all Komondors were u ed in fenced pastures(Tables 1, 2). In all operJ.tions combj.ned, thegreater proportion of lamb mortalities to coyotes inherds guarded by .-\kbash compared to GreatPyrenees likely was related to relatively greater useof.-\kbash on open rJ.nge and the associated greatersheep mortalities there rather than Akbash being

tGuarding-dog beha iors Number Very

producers important

\ ery aggressive to predators 58 88\'erv attentive 10 sheep 59 92\'ery trustworthy 59 93Very active/energy level 58 "Fastest for chasing!

irightening predators 59 ,p

\ erv intelligent 57 54

Great Pyrenees and Akbash; however, Anatolianswere rated as effective as Akbash and GreatPyrenees to deter predation; unfortunately, smailsample sizes did not allow enough power to deter­mine significant differences, if present.

Csing li"estock guarding dogs has been success­ful to deter predation on sheep (Green et al. 1984:Coppinger et al 1988; Green and Woodruff 1988:Andelt 1992; Andelt and Hopper, in press).Producers who h:l\·e used 1 breed of dog usuallyhan: r:lted the breeds similar in effectiveness(Green and Woodruff 1988, this study). However.the greater number of producers using muitiplebreeds of dogs that rated Akbash better than GreatPyrenees and Komondors in this study suggests.-\kbash may be the breed of choice in fenced pas­tures :.lnd on rangelands.

.4cknOldedg11lents. I thank the many sheep pro­ducers ,vho responded to this survey. ]. S. Greenprovided suggestions on the survey. S. X Hopperconducted the telephone surveys. :Vl. G. Fuentesobtained telephone numbers of producers andentered the data. K. P. Burnham prOVided st:ltisticaladvice. P.:\. Vohs reviewed an earlier draft at' themanuscript.

literature cited.'--'OELT. \\: E 1992. Effecti",:ness of H\'e5lOck gUJrding Jog' I<Jr

reducing predation on domestic ';heep_ "'ildlife ~ocjerv

Builetin 20:55-62.

.\."OELT. 'X: E..\."0 S. :\. HOPPER. In press. Live,tOck guud Jogs

reduce predation on domeslic sheep in Colorado..Journal Il["Range .\lanagemem.

COPPI."GER. R.. L. COPI'I:"GER. G. L..."GEL H. L. GETTLER..ISO J. LuRE'Z.

1988. .-\ decade of use oi Ih'eslOck guarding dogs.Proceedings Veflebrale Pest Conference 13:209- 21-i.

COPPI:"GER. R.. ]. LORE:"Z....."0 L. COI'PI:"GER. 19R3a. Imroducinglive rock guarding dogs to sheep and goal producers.

Proceedings E:15tern Wildlife Damage Control Conierence1: 129-132.

COI'PI:"GER. R.. j. LORL"z,]. GI.E."OI:'l:'lI:"G..1.'-"0 P. PI:".\RDI. 19R3b.Anentiveness of gUJrding dogs for reducing predJtion ondomeslic ~heep. journal of Range .\Ianagement 36:T5-279.

CREE:".]. ·...\.'\"o R. .-\. WOOORCFF. 1983. The use of Eunsian dogs

to protect heep from predalors in :\onh AmericJ: J -umma·r~' nf research :II the U.S. Sheep Experiment Slalion.

Proceedings 81.>tern Wildlifc Damage Comrol Confercnce1:11'J-12-l.

GRJ:ES.]. S., .\.'D R. A. WOODIII FE 19SH. I3recd comparisons andcharJClcrbtics of usc of livestock gUJrding dogs. Journal of

R;mge .\bnagement -i I :2'19-251.

GREES.]. S., \'0 R. A. WOOORt FE 19H9. Livcslock-gu:lrding dogsreduce depredalion by be;lrs. P:lgcs ~9-5-i ill \1. Bromley,

editor. BeJr- people conllicls: proceeuings IJf a sympOSiumon mJn:lgement ,Ir.llegie·. :"onllwe,1 Terrilories

Depanment of Renewable Resources. Yellowknife.

GREE:".). S...':"lJ R..-\. WOt)I)RI fF. 19<)0..-\DC guarding dog programupd:lte: a fOClLs on managing dogs. Proceeding, \'eflebrale

Pesl Conference 1';:235-_36.

GREE:".]. ' .. R..-\. WOOURI FF..'.'U T. T. TIELI.ER. 198-4.Livestock·guarding dogs for predawr cOlllrol: COSIS. beneliLs.

and practicalilY. Wildlifc Society Bulletin 12:-i-i-SO.

U:"H.I.RT. . B.. R. T. STER:"EII. T. C. C.\~RJ(:\:". 1..'1) D. R. HE' 'E. 19-9.KOIllondor guard Jogs reduce ,heep losses to coymes: a

preliminary e\-:llu:llion. .Inurn:ll nf Rang" .\1:Jnagement

:'2:238-2-t 1.\kGRE\X·. J. C. 1.:\0 C. :,_ UI.\KL'I.EY. 19H2. How Komondnr dogs

reduce ,;heep lo>ses lU CO\·OIes. journ;d Ili Range

.\Ianagement .35:(,9.:\-(,9(,.

.\IILLlKE:". G..-\_. "I) D. E..Iol'-"'o,. 1<J:-;-i ..-\.I1:1I1.·,is Ilf l11e,,~' uala.Volume I: Designed experiments. Lifelime Le:lrning

Publications. l3elmonl. CJlifi,rniJ.'\.\TIO:\.\L.\(,RICl L11. R.\I. :,nmTIC> ~I.R\ ICE. 1<)I)S. ~heep .lIld Iamb

denh loss J1)9-i. L'niled SUles D"p:lflJl1el1l of .\griculture.

:\alional .-\!tFicullur:ll ~lall'lics ~epice. '\.-\~S ~t:lIT RepoflLDP :-10. <)';·01. \\·a,hingwn. D.C.

:'.-\S [:",nn'TE 1,<:. 19HH. SAS/ST.-\T I',er S guide. ~elc.::lse 1>.0.3 edi­

lion. SAS InslitUte Inc.. C:lP·. :'\nnh C:lrolina.

S\\lU-E. D.). 19'JO..\lullipJc cnmpar;,;on proceJures. the prJcti­cJ..l Solulion..\merion ~Ialblician -i-i: 1-~-lH().

William F. (Bill) Andelt (right, wilh :\kbash gUdrding dog and asheep producer) is an aS50ciate proiessor and extension wildliiespeciali5t in the Department oi Fi5hery and Wildlife Biology atColorado State LJniver5itv. He received hi5 B.S. 1197--1) ,JIld M.S.·19761 degrees in wildlife science5 irom the Univer5ity or--':ebraska Jnd his Ph.D. f19821 in zoology rrom ColorJdo SlateL'niversitv. Bill's interests include research Jnd exten­sion related to managing humJn-wildlire conrlicts and'he ecology or carnivores.

Associate Editor: Conner