Bear Management and Sheep Husbandry in Norway, with a Discussion of Predatory Behavior Significant for Evaluation of Livestock Losses

BEAR MANAGEMENT AND SHEEP HUSBANDRY IN NORWAY, WITH A DISCUSSION OF PREDATORY BEHAVIOR SIGNIFICANT FOR EVALUATION OF LIVESTOCK LOSSES

IVAR MYSTERUD, Zoological Institute, University of Oslo, Postbox 1050, Blindern, Oslo 3, Norway

Abstract:During the 19th centurythe brown bear (Ursus arctos) populationin Norway was reducedto remnantlevel. The populationhas since been restored and recently seems to be increasing. Concern is present for bear managementin connection with sheep predation, as sheep husbandryis importantthroughout Norway, the stock in 1976 amountingto 1.6 million animals. The managementtechnique now practiced combines selective hunting of troublemakerswith monetarycompensation for sheep killed. The numberof sheep killed by bears is insignificant comparedwith the total sheep mortality, and bear predationis importantonly locally, primarilyin areas in Hedmark,Hordaland, and Finnmarkcounties. Ethical argumentsagainst bears are raised in connectionwith observationsof overkill, and a researchprogram has been initiated to analyze predationpatterns in greaterdetail. Overkill by bears is not restrictedto surpluskilling. In most cases, small amountsare consumed from each carcass - nutritionallyvaluable parts such as breast fat deposits and udders. This behavior may representextreme food selection underplentiful prey conditions and should be comparedwith selective grazing among herbivores.The organizationof behaviorin predatorymammals relevant to livestock losses is discussed.

Several grazingtechniques and herdingsystems have mentingon the manuscript;and to L. Wold for improv- been adaptedto the Scandinavianforest and mountain ing the English. The Hedmark Bear Research Pro- ecosystems and are typical of vast areas covered with gram is supportedby the Norwegian Research Council morainedeposits and mineralsoils. During the Middle for Science and the Humanities, FridtjofNansen's and Ages, highly differentiatedchalet systems were already affiliated Funds for the Advancement of Science and organized in Norway, keeping and rotating herds of the Humanities, and the Directorate for Wildlife and grazing animals under the protection of herdsmen far FreshwaterFish. out in wilderness areas. PRESENT STATUS Predation by the brown bear and other large carni- OF SHEEP FARMING vores representsa traditionalconflict in Norway. Bear Commercial sheep husbandry in Norway has re- and sheep may compete for space and plant biomass in mained small in scale, and even today small herds of outlying grazing areas, and the bear can become an sheep are kept mostly by individual farmers. Flocks of efficient on predator sheep. Large predatorshave cer- sheep are introduced without herdsmen into grazing been tainly influential in the development of the chalet lands, many of which are topographicallyrough, and culturein Scandinavia;danger of predationwas an im- most flocks are tended at rather irregular intervals. consideration portant when choosing localities for new Total stock figures show the development of sheep chalet farms. Reinton in (1955), studies of the chalet farmingin Norway from 1940 to 1976 (Fig. 1). During culture in claimed that Norway, many chalet barns the war the number was kept low and figures ranges were originally built equally as defense against bears and as from inclement weather. protection Dispropor- 23 - tionately strong doors and locks were distinctive fea- tures of such buildings. The practice of allowing ani- 221 2 o mals to at z graze night, periodically common on the 0 1g- European continent, was seldom practiced in Scan- in ~s 17- dinava, due, in part, to the danger of predation(Szab6 uJ I6- 1970). Some claim that the earlier in rI herding system uqa.Ln LL Norway was originally developed for the sole purpose 0 of protectinglivestock from predators.Most herdingof sr 12 - livestock :D ceased in Norway when the numbersof pred- z atorsgradually decreased in the outlying grazing areas. 08 - The author is indebted to P. Bendixen, Directorate for Wildlife and Freshwater Fish, for valuable ~ help 1940 41 42 43 44 45 46 47 48 49 50 5 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 with preparationof the tables; to A. Seierstad and N. YEAR of Stenseth, University Oslo, for reading and com- Fig. 1. Totalnumbers of sheep present in Norwayon 20 June each year, 1940-76. 234 BEARS THEIR BIOLOGY AND MANAGEMENT between 0.8 and 1.0 million. Postwar, however, the stock increased to 1.7 million, and since then the w 1800 - 300 _j ---US DOLLARS - NORW.CROWNS number has fluctuated between 1.6 and 2.1 million, y z 1600 with peaks during 1950-55 and 1964-68. In the 1970s, o 1400 the stock remainedat about 1.6 million animals, but is wL m 1200 200 m" now increasing. The 1976 total was 1,667,488, - 1000 _i whereasthe 1974 and 1975 figures were 1,632,217 and z o 1,639,313, respectively, an increase of 2 percentin the 2 LO800 latter < 600 100 > years. Z C At agricultural experts advocate the >- 400 present, 2 m o economic policy of increasing sheep production in ( 200 Norway. A prognosis for 1990 predicts 1,993,000 I I I I I I I I I I 1 animals. One of the national is to reduce the 1966 67 68 69 70 71 72 73 74 75 76 goals YEARS importationof certain meat products and fodder, thus becoming more self-sufficient throughbetter utilization Fig. 2. Economic compensation awarded per sheep for ewes killed by brown of national resources. There are few ecological or bears in Hedmark County, 1966-76. Amounts given are in Norwegian crowns a of (solid line). U.S. dollar equivalents (stippled line) are based on current exchange biological argumentsagainst strategy self-support, values. which includes increasinguse of local outlying grazing bears. 2 some of calcu- areas above the present level of sheep production. killed by Fig. gives examples lated prices per sheep for ewes killed in Hedmark LOSSES AND COMPENSATIONS county in recent years. One way to influence sheep producersin a positive losses for as beardepredations in Sheep compensated direction has thus been to adequatelycompensate their 1966-75 are shown in Table 1. As can Norway during economic losses. A futurestrategy could be to increase be seen, the yearly compensated loss fluctuated be- this compensationto profit level. tween 30 and 240 animals. The losses are presently to the total stock. losses small in relation sheep Sheep STATUS OF THE BEAR POPULATIONS are, however, generally restricted to certain small areas, the most importantfound in Finnmark (Wikan From an original population consisting of several 1970, 1972), Hordaland (Elgmork and Mysterud thousandanimals, the brown bear was relentlessly shot 1976), and Hedmark (Myrberget 1968; Mysterud in Norway during the 18th and 19th centuries and re- 1974, 1975a, 1975b) counties. Some losses also occur duced to a remnantlevel. The approximateannual kill in Opplandand Buskerud(Elgmork 1976b), SBrTr6n- around 1850 averaged225 individuals. Hunting statis- 1850 to delag, Nordland, and Troms counties. Economic tics indicate a steady populationdecline from the compensation, which does not representactual market 1925 throughoutthe country(Myrberget 1969). By value, is awardedfor sheep that can be documentedas beginning of the 20th century, the populationhad been

1966-75. Table 1. Numbers of sheep compensated for as being killed by bears, by county, in Norway,

Counties Total Year Hordaland Nord Nordland Troms Finnmark Hedmark Oppland Buskerud TrndelagTri6ndelag Trn6ndelag

237 1966 39 4 4 0 190 0 0 0 0 O 78 95 1967 12 5 0 0 0 0 0 0 168 1968 40 6 0 0 0 0 0 0 122 75 1969 0 0 17 0 0 0 0 0 58 26 35 1970 0 6 3 0 0 0 0 0 51 69 1971 0 0 18 0 0 0 0 0 51 1972 0 0 19 0 32 0 0 0 0 87 1973 63 0 21 0 0 0 0 3 0 180 1974 32 0 0 31 0 0 117 0 0 87 222 1975 48 5 0 69 0 13 0 0 422 Total 234 26 82 100 222 13 117 3 1,219 BEARS AND SHEEP * Mvsterld 235 greatly reduced, and between 1920 and 1930 it was at a removal to solve the problem of stock-killers. The minimum. In about 1940, Olstad (1945) believed that technique now being tested includes shooting only residentpopulations were present only in the Vassfaret troublemakersthrough ordinaryhunting - careful re- area in Oppland and Buskerud counties (Elgmork moval of predatorybears at the time and place of the 1976a) and in South-Varangerin Finnmark.However, problem, preferably only the individual actually in- recent investigationshave shown that remnantpopula- volved. The removal of troublemakersis undertakenby tions have also survived in other areas (Mysterud small groups of hunters from the local areas who are 1977). Myrberget (1969) estimated the population at paid for their efforts by the federal government. The 25-50 animals in 1969, but it must now be roughly personnelof each huntinggroup is selected by the local estimated at more than 100 individuals. Decreasing game managementunit and approved by the wildlife trends have changed during the latter half of the 20th administration.The hunters are not professionals, but century, and the populationsnow seem to have entered hunt only when stock-killersappear in their area. a period of increase. The total numberof bears killed In 1968, 1969, and 1970, permits to remove stock- has remained at a stable level since 1940; the distribu- killers duringthe protectedperiod priorto 15 June were tion of the kill by counties is shown in Table 2. Since given to S,Br-Varanger municipality, Finnmark

Table 2. Hunting statistics for brown bears in Norway, 1940-76. (Based partly on Myberget 1969.)

Counties

Years Total Hedmark Oppland Buskerud Telemark Sogn og Nord- Nordland Troms Finnmark Fjordane Trg6ndelagljr- Trj6ndelag

1940-49 2 1 4 1 1 0 2 4 0 1 16 1950-59 4 0 1 0O 1 0 2 10 0 5 23 1960-69 5 0 0 0 0 1 1 3 0 11 21 1970-76 2 0 0 0 0 0 1 0 2 6 11 Total 13 1 5 1 2 1 6 17 2 23 71

1970, 11 bears have been removed from the popula- County, and 1 bear was shot in 1968. During 1973, tions. Methods of removal include legal hunting (3), total protection was enacted and the first permit to re- shooting of stock-killers (4), self-defense killing (1), move a stock-killing bear was issued in Trysil, Hed- and poaching (3). mark County, where a subadult male was shot (Mys- terud 1975b). In 1975, Trysil municipalitywas given a PROTECTION AND MANAGEMENTOF BEARS new permit, but no bear was killed. In 1976, 11 permits Since 1940, measuresto protectthe brown bear have to remove 8 different bears were issued in 8 mu- been as follows: During 1940-41, the species was pro- nicipalities in Hedmark (8), SA6r-Tr,6ndelag(2), and tected during the denning period from 1 November to Finnmark(1) counties. Seventeen huntinggroups were 15 May. During World War II, 1942-45, bears were formed, 2 in municipalitieswhich did not receive per- given total protection.After the war, the protectionwas mits. One subadultmale was shot in Trysil municipal- removed except for the period 1 November-15 May. In ity by the collective efforts of the 15 hunting groups 1966, this protectionwas extended to 15 June in Troms that were in operation. Thus, only 3 bears have been and Finnmark counties, primarily to prevent hunting removed by authorized hunters during the period with snowscooters (Myrberget 1969). On 28 May 1968-76; this low efficiency has led to a discussion on 1971, the bear was given total protection in Oppland differentbear-hunting techniques. Removal of specific and Buskerudcounties, but not until 11 May 1973 was killers is considered the most selective and efficient the species given total protectionthroughout the coun- means of contending with the predatorproblem (Cain try. It is legal, however, to kill bears attackinghumans et al. 1972), and it will be importantto develop an and livestock, regardlessof protection laws. efficient strategy for future removal of troublemakers. A significant political pressurefor removal of bears In spite of the acknowledged hazards to livestock, stills emanatesfrom interestsconnected with sheep and opinion polls, even from counties with the greatest reindeer (Rangifer tarandus) husbandry. The federal losses of sheep, show that a majority are in favor of wildlife administrationis attempting to use selective preservinga bear population. 236 BEARS - THEIR BIOLOGYAND MANAGEMENT

SHEEP-KILLINGPATTERNS OBSERVED IN behaviorof a predatorthat kills withouteither itself, its BEARS young, or membersof its social unit attemptingto con- sume the prey, although they have free access to the Much more data are needed from detailed studies of carcass and the prey is among their potential food which individualbears develop into stock-killersbefore species (Kruuk 1972). we can evaluate the long-term consequences of their Surplus killing occurs primarilyin closed, artificial removal from the population.Hedmark County, where conditions, for instance when a fox (Vulpes sp.) gets the bear populationhas long been notorious for killing into a henhouse or mink (Mustela vison) into a hatch- sheep and where some damage to sheep flocks is more ery pond. Such events are sometimes describedas kill- or less constant (Mysterud 1975a), has been selected ing "orgies," a phenomenon exceptionally detrimen- for studies, and data on sheep-killing bears have been tal to domestic livestock if they have no means of es- collected since 1968 (Mysterud 1974, 1975b, unpub- cape (Fox 1971, Nesse et al. 1976). Young and lished data). The investigationsso far have shown that Goldman (1946) mention a case where a puma (Felis sheep predation incidents in the area can be roughly concolor) entered a sheep pen and, in the course of 1 grouped into 3 categories: night, killed 192 sheep. Blocking of escape routes by fences may be an importantfactor in explaining surplus (1) Sheep killed by adult male bears and used as a killing, but it also occurs in unfenced naturalecosys- regular food source by the male and/or associated tems. The phenomenon applies to mammalian pred- resident on or near sheep grazing pas- family group ators in general; under conditions where normal anti- tures. predatorreactions of the prey are blocked, surpluskill- (2) Sheep killed, often in numbers, by subadult ing seems to be released. Among bears, several exam- males that have established home ranges on the ples of surpluskilling have been describedfor the polar areas. Such individuals may develop into grazing bear (Ursus maritimus). Even Nansen (1924) called habitualsheep-killers and troublemakers. attention to the fact that polar bears killed far more (3) Occasional sheep killed by stragglingor transient young seals than they consumed (L6n, 1970). Perry bears, most often of little significantto the total kill. (1966) mentioned 1 case in which a polarbear killed 21 Field observationof predationpatterns in the study narwhals (Monedon monoceros) enclosed by ice in a area has shown that the usual patternof bear predation small pond, and Freeman(1973) reportedmultiple kills is that 1 or a few sheep are killed at a time and are either of beluga (Delphinapterusleveas) by a solitary bear. consumedor cached for futureconsumption. However, During field investigations into high densities of situationscan arise in the outlying grazing areas where small rodents, large numbers of lemmings (Lemmus a numberof sheep are killed within a relatively limited lemmus) and other small mammals such as shrews are time and space. In 1 case from Hedmark, 26 of a herd often bitten to death but left uneaten(Mysterud unpub- of 28 sheep were killed during 1 night. Other overkill lished data). One attempted explanation of this has incidents occurredin 1973 and 1976 (Mysterud 1975b, been that lemmings, for example, contain substances unpublisheddata). The numberof sheep struckdown by that make them less coveted as prey (Hagen 1952, An- the predator,in some cases clearly exceeding the im- dersson 1976). However, underother conditions, large mediateor short-termneed for prey biomass, and multi- quantitiesof lemmings are consumed by the same pre- ple kills and unconsumedcarcasses seem to be reported dator. in almost every study of predatorlosses (Rowley 1970, During certain winter conditions, prey species may Henne 1975, Dorrance and Roy 1976, Nesse et al. trampthrough snowpack while a predatorwalks upon 1976). Little seems to be known about the biological it. Surface structure of snowpack seems to be ex- significance of such killing patterns. This observed tremely importantin some predator-preyrelationships; phenomenonwill thereforebe exploredand explained in and on certain types of crusts, examples are known a theoretical discussion of the general organizationof where brown bears have killed a number of moose mammalianpredatory behavior. (Alces alces) (Hellgren 1967, Wikan 1970). However, the carcasseswere not left completelyuneaten; the pred- SURPLUS KILLING ation thereforerepresents killing patternno. 2 discussed One well-known type of overkill reported among below. mammalianpredators is surpluskilling of prey (Kruuk The caribou-huntingbehavior of American Indians 1972, Curio 1976). Surplus killing is defined as the and Eskimoes, summarized by Kelsall (1968), may BEARS AND SHEEP * Mvsterud 237 also be explained as surpluskilling. During certain in- of the udderare removed (Fig. 3). These incidents may cidents, when natives are faced with large numbersof representextreme food selection as defined by Stenseth caribouand have plentiful ammunition,they frequently and Hansson (1977): a selection of small but nutrition- seem to go berserk and fire blindly into the animals ally valuable parts of the animal. Such killing is also until ammunition or caribou are gone (Kelsall part of human hunting behavior, exemplified by the 1968:221). killing of caribou by northernIndians for removal of Surplus killing has been assumed to occur only in tongues (Kelsall 1968). High sheep density may well extremely rare cases (Curio 1976). This is probably a representan ecological situationwhere individualbears correct assumption when comparing the surplus-kill develop optimal eating habits - "grazing" on sheep percentagewith the usual consumptionkilling. When it udders and breast fat deposits. This preferenceconsti- does occur, however, it may be of quantitativeimpor- tutes a combination of high-energy fat storage and tance to the group or populationaffected. Dorranceand udder tissue rich in proteins, minerals, vitamins, etc., Roy (1976) have, during their investigations on sheep and representsa nutritionallybalanced intake. If this losses to predatorsin Alberta, Canada, reportedthat "organ grazing" constitutes extreme food selection, the probability of predation was lower in confined whole carcasses should not be consideredas units when flocks - those kept in pens or sheds throughoutthe a predator'seating behavior is considered. This eating year- than in free-rangingflocks. But when predation patternmay be better comparedwith selective grazing did occur in confined flocks, it was particularlysevere. among herbivoresthat remove specific parts of plants. As a result, predationlosses were highest in confined Field studies of a variety of species show that herbi- flocks and lowest in range flocks. Conclusions con- vores usually select plant items so that the net gain of cerning the ecological significance of this behavior metabolizableenergy is maximized, as shown by Sten- should thereforenot be drawn until field investigations seth at al. (1977) for the field vole (Microtus agrestis). provide information on the frequency of this form of The breadthof the food niche of field voles has been killing. shown to vary with the density and quality of all avail- EXTREME FOOD-SELECTION KILLING able food items. With the more extreme organ selection and smaller amount consumed a Overkill phenomena other than true surplus killing, per sheep, larger numberof animals must be killed to meet a bear's food where no part of the animal would be eaten (Kruuk 1972), have been observed. Sheep carcasses investi- requirements. It is not the of gated in most overkill incidents in Hedmark are not purpose this paperto presenta detailed model of this behavior, but it is not difficult to theoreti- without small eating marks (Mysterud unpublished describe such data). In such cases, the breast of the animal is opened cally biological phenomena by simple, continuousfunctions, here in 3 in a stereotypedway, andbreast fat deposits and/orparts presented steps. When disregarding certain factors, the average amountconsumed from each carcass may be described as a function of sheep populationdensity. The number of animals killed per unit time (z) multiplied by the amount, measuredin weight (v), consumed from each carcass satisfied the food requirementof the bear. If this requirementis a rough constant, K, and the prey animalscan be regardedas "homogeneous food," then K zv = K. Hence, v = -, which expresses a hyperbola. Obviously the homogeneity assumption is an over- simplification, but one might reasonably present the function v(z), shown in Fig. 4. The numberof animals killed per unit time (z) may be assumed to be an increasing function of the sheep Fig. 3. A ewe killed by a raiding bear in Hedmark, southern Norway, 1976, population density (x). Such a functional response showing the consumption marks typical of extreme food selection on sheep. curve, modified from is The bear has cut the hide, the and consumed the fat Holling (1959), denoted by exposed breast, deposit and it can have 1 together with small amounts of udder. Instead of eating more from this carcass, z(x), of several forms. In Fig. 5, an other nearby ewes were killed and the selective eating pattern repeated. arbitraryillustration of such a function is shown. If we /

238 BEARS - THEIR BIOLOGY AND MANAGEMENT

> A sity, bearsearch, hunt, chase, andcapture efforts are, of ^ Vtot- CLASSIC CONSUMPTION u course, all minimized. Let us explore in some detail the succession of ele- I ments organized as part of the predatorybehavior se- relevant to an of this last state- r - quence understanding O0 ment. LL _ 0D THE PREDATORY BEHAVIORSEQUENCE r> \ INCREASED FOOD SELECTION L) z Restricteddefinitions of the of in 0 concept predation z A D3 EXTREMEFOOD SELECTION > o SURPLUS KILLING m) Vtot <~ I I I I I I I I I , I U 0 ANIMALS KILLED PER UNIT TIME z U I< - u Fig. 4. One basic relationship in predation incidents on sheep is the amount eaten from each carcass(v) as a function of the number of animals killed per unit O time (z). The function runs from classic consumption, where 1 animal is killed L and completelyconsumed (z = 1, v = v ,O, to surplus killing,where manyanimals are killed which are not consumed (z -- x, v = 0). Between these values, increased selection of parts develops with increasing z, a phenomenon U prey zZ denoted as extreme food selection (z - x, v - 0). 0 \ 0 Z

Lu/

SHEEP POPULATION DENSITY x

Fig. 6. Amount eaten from each carcass (v) as a function of sheep population - - The function is is the number of animals w z(x) density(x). expressed asv(z(x)), wherez killed per unit time (cfr. Figs. 4, 5). The function v(z(x)) can have several forms; 1 is here. _j only example given

--- ?z -

SHEEP POPULATION DENSITY x

Fig. 5. Functional response curve z(x) for bear-killed sheep, wherex denotes sheep population density and z is the number of animals killed per unit time. The function z(x) can have several forms; an arbitrary example is given here. combine the curves in Figs. 4 and 5, we see that the weight or precentageconsumed from each carcass be- comes a decreasing function of x (Fig. 6). The higher population density (x) means more animals killed per unit time (z) and a smaller amount (v) eaten from each carcass. This function, denoted by v(z(x)), can have several forms, of course; only 1 example is illustrated in Fig. 6. It must be pointed out that the z(v) functions actually depend on several other factors, not included as argumentsin the function. Of special interest is the relative ease with which a bear can capture a sheep. This factor perhaps stem from some sort of optimiza- 7. of behavioral associated with tion between the requiredto eat and digest most Fig. Organization specific patterns predation by energy large mammals. Loop A describes classic killing-and-consumption predation, B or all of an animal, comparedto the energy expendedin and C surplus killing, and D extreme food-selection killing. All patterns have capturingan animal. On pastureswith high sheep den- been observed in areas where brown bears prey on sheep. BEARS AND SHEEP * Mysterud 239 most textbooks generally include the killing and eating andother enclosures, the release of such a sequencemay of animals (Hassel 1976), but Curio (1976) has argued be explained simply by blockage of the normal flight that "eating" should be avoided. Curio (1976:1) de- reactionof the prey. The predatorbecomes linkedup in a fined predation as "a process by which an animal continual C-loop: As one animal is killed, a new spends some effort to locate a live prey and, in addi- stimulusto chase is immediatelyreceived. The stimula- tion, spends anothereffort to mutilate or kill it." For tion to stop huntingis difficult to isolate (Grant 1972). our purpose it is advantageousto furtherbreak down Extremefood selection might similarly be described the definition. In the simplified model presentedin Fig. by the sequencechase -> capture-> kill -> eat -- chase 7, separatebehavioral elements, each assumed to have -> capture ->..., with only small, nutritionallyvalu- its specific threshold motivation, have been identified able parts eaten after each kill (Fig. 7, loop D). in the organization of predatorybehavior. All of the That one or several behaviorscan be smoothly linked observed patterns of predation associated with large together is a matterof efficient adaptationin different carnivorescan be explained by this model. species when thresholdsare low. Even the capture be- Classic killing-and-consumptionpredation follows havior in the model is sometimes observed separately the sequence shown in loop A of Fig. 7. Hunger pro- from killing, for instance when polar bears play and vokes search for prey, and the hunt is then initiated. throw young seals in the air (Perry 1966) or young cats The predator may locate fresh scent trails (olfactory capturemice and play with them without killing them. stimulus), or may hear (auditory stimulus) or sight In some cases, extreme organ selection has been ob- (visual stimulus) the prey. Different strategies have served on sheep which have not actually been killed been developed for approachingor surprisingthe prey. (Mysterud unpublished data). The survival of such In certainspecies, the location of prey culminatesin the animals may be explained as shock-bite recoveries predatorchasing the prey in the attemptto capture it. (Mysterud 1975b), general shock recoveries, incom- Specific behavior patternshave been developed in all plete killing due to humandisturbance, incomplete kill- species to capture, kill, andeat the prey. Rest is usually ing behavior by young animals, or eating performed induced or associated with a period of satiation, and withoutkilling. From purely biological arguments,it is after assimilationof the food, hungeris again induced. not unlikely that, undercertain circumstances, handling Search -> hunt --> capture -> kill -> eat is thus an of prey animalsbecomes so easy thata behavioralsequ- idealized and simplified behavior sequence, almost ence following a modified D-loop, chase - capture-> dogmaticallydescribed for most mammalianpredators. eat -> chase ... (Fig. 7), could develop where thereis a However, Kruuk (1972) has observed in the African net gain of energy by omitting killing. There is also the wild dog (Lycaon pictus) that both satiatedand resting possibility thatprey-killing may have evolved as partof animals can be stimulatedto hunt anew by the sudden the eating process and that specific killing bites or appearanceof prey near them. A chase will then release methods are lacking. Henne (1975) has reported2 in- the componentscapture andkill, but not necessarilyeat stances in which coyotes fed on sheep before they died, (Fig. 7, loop B). Thus, chasing seems to be motivated and the same has been observed with coyote predation differently from killing and eating (Kruuk 1972). Fox on mule deer (Odocoileus hemionus) (Cahalane 1947) (1971) suggested thatthe prey-killingresponse of canids and elk (Cervus canadensis) (Robinson 1952). A re- has a very high satiationlevel and that prey-killingcan stricted concept of predationcould be used synonym- therefore continue beyond the nutritionalneeds of an ously with killing behaviordefined, for instance, as the individualsimply because the prey is available and vul- removalof live prey animalsfrom populations.But even nerable. However, another study of sheep-killing by this definition will not be specific, as killing is also an coyotes (Canis latrans) under captive conditions element of intraspecificcompetition (Curio 1976). showed that food deprivationhad no descernibleeffect on the killing behaviorbut did influence feeding activity DIFFERENTIALRATE IN MAMMALIAN on kills. The observationsdirectly suggested that hunger PREDATION is not always the primarymotivation for predatorybehavior (Connolly et al. 1976). The predatory strategies of mammals may be or- During surplus killing incidents, it seems that en- ganized so that removal of prey animals can be de- vironmentalconditions can release behavior sequences scribed at different rates, dependenton environmental -> of the type chase capture -> kill -> chase -> capture factors relating to escape behavior, etc. Consequently, -> kill -> chase ->... (Fig. 7, loop C). In fenced areas mammalianpredator pressures may be far more effi- 240 BEARS THEIR BIOLOGYAND MANAGEMENT cient than supposed from estimates of whole carcass (Cringanand Harris 1972). Predationpressure exerted consumption, and the speed at which prey animals are by large carnivores may be of prime importance in eliminated may be maximized under various cir- developmentof role play, space use, and social evolu- cumstances. Naturally this may be of significance in tion, in both prey and predatorspecies, expressed by both artificial and naturalsituations where the popula- Edmunds(1974:14) as an "armedrace" between pred- tion of prey animals is excessive and the number of ator and prey, both of which constantly develop new predatorstoo few to control prey populations through behavioral patterns involving both gene pools. Prey eating procedures. In most situations, stability of prey removal rates in husbandry - differential livestock populationscan probablybe maintainedin partthrough loss sizes - are an importantfield of study in evalua- the pressureexerted by maintainingclassic consumption tion of these evolutionaryconsequences. killing. The most importanteffect of the mechanismof For livestock productionit is basic to realize (1) that differential rate operating through extreme food- evolutionarybehavioral equilibria in open systems exist selection and surpluskilling may be in systems where a before the introductionof flocks of animals whose be- large surplusof cyclic animalsregularly must be killed havior or social structureis not adaptedto the prevail- off, which is usually the case in large areas of the ing ecosystem. Sheep and domestic reindeer, the 2 northernhemisphere. This mechanismis also important most important rangeland species in Norway, often for any population that undergoes irregular fluctua- graze in socially unstructuredaggregates and periodi- tions. cally graze within fenced areas, making them highly In the construction of models describing predator- vulnerable.(2) The majorityof domestic animalspecies prey relationships,differential predation rates of these have greatly reduced locomotive capacity compared types will complicate the descriptionof predationeffi- with wild animals. Productionof sheep is today guided ciency due to erroneous estimates of prey removal by commercialaims such as quantityand quality of wool drawn from energy arguments, which are most often and meat, and natural defense and/or escape based on consumptionof the whole prey animal. That mechanismshave been generally bred out in exchange is, when calculating the needed prey numberbased on for increasedproductivity or increasedease of managing energy and nutritionaldemand per individual predator the animals (Howard 1974, Klebenow and McAdoo and multiplying by the estimated size of the predator 1976). More detailed knowledge of predatorbehavior population, there is strong evidence that prey removal should bring parametersof prey behavior patterns, so- might be underestimated.From investigationsof bears cial structure,locomotor capacity, etc., withinthe scope and sheep in Hedmark, it is already clear that such of livestock researchto minimize losses. simple evaluations cannot be used in connection with Ecosystem managementwhich aims at maintaining sheep-bearrelationships. predator and livestock populations within the same space, therefore, needs a knowledge of predator-prey WHAT ARE THE FUTURE RESEARCH NEEDS? relationships for development of good predatorman- agement programsand estimation of optimal livestock In regardto predationand livestock, the only conclu- densities and tolerable loss sizes. However, an under- sion one can draw from literatureis that much more standing of the predatorbehavior sequence and live- research is needed (Jobman 1972). This discussion stock removal rates ought to be regardedas of utmost proposes that predatorybehavior as a strong selective importance, whether we are to introduce more sheep force and control mechanism for surplus population into the habitat of a small bear population or into the numbers might be more biologically significant than habitat of an increasing bear population. At present, tissue-eating and the associated transferof matter and these 2 alternatives seem to be the only realistic choices energy in ecosystems emphasized in energetic contexts in Norway.

LITERATURECITED CAINE, S. A., J. A. KADLEC,D. L. ALLEN, R. A. COOLEY, M. G. HORNOCKER, A. S. LEOPOLD, AND F. H. cause of ANDERSSON,M. 1976. Lemmus lemmus: a possible WAGNER. 1972. Predator control -1971. Univ. of coloration and behavior. J. Mammal. aposematic Michigan Inst. Envir. Qual. Rep. 207pp. 57(3):461-469. CONNOLLY,G. E., R. M. TIMM, W. E. HOWARD,AND W. CAHALANE,V. H. 1947. A deer-coyote episode. J. Mammal. M. LONGHURST.1976. Sheep killing behavior of captive 28(1):36-39. coyotes. J. Wildl. Manage. 40(3):400-407. BEARS AND SHEEP * Mysterud 241

CRINGAN,A. T., AND L. D. HARRIS,eds. 1972 Predation and MYRBERGET.S. 1968. Bj6rn og sau i Hedmark. (Bears and its management.Colorado State Univ. Grad. Sem. Pap. sheep in Hedmark.) Naturen 92(2):67-80. (In Norwe- 132pp. (Mimeogr.) gian.) CURIO,E. 1976. The ethology of predation.Springer, Berlin. . 1969. Den norskebestand av bj6rn, Ursus arctos L. 250pp. (The Norwegianpopulation of brownbear, Ursus arctos DORRANCE, M. J., AND ROY. 1976. Predation losses of L.) Meddr. Stat. Vilt. Ser. 2(29):1-21. domestic sheep in Alberta. J. Range Manage. MYSTERUD,I. 1974. Sau drept av bj6rn, Ursus arctos, i 29(6):257-260. Engerdal, Syd-Norge, i arene 1953-1963. (Sheep killed EDMUNDS,M. 1974. Defence in animals. Longman, Essex. by brown bears, Ursus arctos, in Engerdal, south Nor- 357pp. way, during the years 1953-1963). Fauna, Oslo ELGMORK,K. 1976a. A remnant brown bear population in 27(3):121-138. (In Norwegian, English summary.) southern Norway and problems of its conservation. Pages .1975a. Bj6rneni Hedmark.Bestand, slagbj6rnskader 281-297 in M. R. Pelton, J. W. Lentfer, and G. E. Folk, og forvaltning. (The bear in Hedmark. Population, Jr., eds. Bears - their biology and management. IUCN stock-killing damage and management.) Jakt-fiske- Publ. New Ser. 40. friluftsliv 104(9):16-21, 48. (In Norwegian.) 1976b. Sau og bj,rn i Vassfartraktene1969-73. . 1957b. Sheep killing and feeding behaviour of the (Sheep and bear in the Vassfaret area 1969-73). Norsk brownbear (Ursus arctos) in Trysil, south Norway 1973. Natur 12(12):56-58. (In Norwegian.) Norw. J. Zool. 23(4):243-260. , AND I. MYSTERUD.1976. Bjrn i Hordaland. (Bears .1977. Problemeri forskningog forvaltningav bjrn i in Hordaland.) Jakt-fiske-friluftsliv 105(8):44-46. (In Norge. (Problems in researchand managementof bears Norwegian.) in Norway.) Viltrapport4:19-51. (In Norwegian, Eng- Fox, M. W. 1971. Behavior of wolves, dogs and related lish summary.) canids. Harperand Row, New York. 220pp. NANSEN, F. 1924. Blant sel og bj6rn. (Among seals and FREEMAN,M. M. R. 1973. PolarBear predationon beluga in bears.) Dybwad, Oslo. 343pp. (In Norwegian.) the CanadianArctic. Arctic 26(2):163-164. NESSE, G. E., W. M. LONGHURST,AND W. E. HOWARD. GRANT,W. E. 1972. Perspectiveson predation.Pages 57-65 1976. Predation and the sheep industry in California, in A. T. Cringanand L. D. Harris,eds. Predationand its 1972-1974. Univ. of California Div. Agric. Sci. Bull. management. Colorado State Univ. Grad. Sem. Pap. 1878. 63pp. 132pp. (Mimeogr.) OLSTAD, 0. 1945. Jaktzoologi. (Hunting zoology.) Cappe- HAGEN,Y. 1952. Rovfugleneog viltpleien. (Birdsof prey and len, Oslo. 249pp. (In Norwegian.) wildlife management.) Gyldendal, Oslo. 603pp. (In PERRY,R. 1966. The world of the polar bear. Cassel, London. Norweigian.) 195pp. HASSEL,M. P. 1976. The dynamicsof competitionand pred- REINTON,L. 1955. Saeterbruketi Noreg. (Dairying in Nor- ation. Studies in Biol. 72. Arnold Publ., London. 68pp. way.) I. Satertypar og driftsformer.Institutt for Sam- HELLGREN,E. 1967. Slagbjmrneni Frostviken. (The killing menlignendeKulturforskning Ser. B, Skr. 48. 48 lpp. (In bear of Frostviken.) Svensk Jakt 105(10):502-504. (In Norwegian.) Swedish.) ROBINSON,W. B. 1952. Some observationson coyote pred- HENNE,D. R. 1975. Domestic sheep mortalityon a western ation in Yellowstone National Park. J. Mammal. Montanaranch. M.S. Thesis. Univ. of Montana, Mis- 33(4):470-476. soula. 53pp. ROWLEY,I. 1970. Lamb predationin Australia:incidence, HOLLING,C. S. 1959. Some characteristicsof simple types of predisposing conditions, and the identification of predationand parasitism.Can. Entomol. 91(6):383-398. wounds. CSIRO Wildl. Res. 15(1):79-123. HOWARD,W. E. 1974. The biology of predator control. STENSETH,N. C., AND L. HANSSON. 1977. Optimal food Addison-Wesley Modulein Biol. 11. Menlo Park, Calif. selection: a graphic model. Am. Nat. (In press). 48pp. , A. MYLLYMAKI,M. ANDERSON, AND J. JOBMAN,W. G. 1972. Predationof carnivores on domestic KATILA. 1977. General models for the population animals. Pages 90-95 in A. T. CRINGAN AND L. D. dynamics of the field vole, Microtus agrestis (L.), in HARRIS,eds. Predationand its management. Colorado central Scnadinavia. Oikos 29:616-642. State Univ. Grad. Sem. Pap. 132pp. (Mimeogr.) SZABO,M. 1970. Herdaroch husdjur.(Herder and husbandry KELSALL,J. P. 1968. The migratorybarren-ground caribou of animals.) Nordiska museets Handlingar73. 389pp. (In Canada Can. Wildl. Serv. Monogr. 3. 340pp. Swedish.) KLEBENOW,D. A., ANDK. MCADOO.1976. Predationon WIKAN,S. 1970. Bj6rn i Sgr-Varanger. (Bears in South Var- domestic sheep in notheasternNevada. J. Range Man- anger.) Fauna, Oslo 23(2):85-101. (In Norwegian, Eng- age. 29(2):96-100. lish summary.) KRUUK,H. 1972. Surplus killing by carnivores. J. Zool. . 1972. Spor som skremmer.(Tracks that scare.) Nor- Lond. 166:233-244. den 76(13-14):420-422, 428. (In Norwegian.) 0. 1970. L,6n6, The polar bear (Ursus maritimus Phipps) in YOUNG, S. P., AND E. A. GOLDMAN. 1946. The puma. the Svalbardarea. Norsk PolarinstituttSkr. 149:1-103. Dover Publ., New York. 358pp.