GRIZZLYBEAR USE OF ARMYCUTWORM MOTHSIN THE YELLOWSTONEECOSYSTEM

STEVENP. FRENCH,Yellowstone GrizzlyFoundation, 581 ExpeditionDrive, Evanston, WY 82930 MARILYNNG. FRENCH,Yellowstone GrizzlyFoundation, 581 ExpeditionDrive, Evanston, WY 82930 RICHARDR. KNIGHT,Interagency Study Team, ForestrySciences Laboratory,Montana State University, Bozeman, MT 59717

Abstract: The ecology of alpine aggregationsof army cutwormmoths ( auxiliaris) and the feeding behavior of grizzly bears (Ursus arctos horribilis) at these areas were studied in the Yellowstone ecosystem from 1988 to 1991. Army cutwormmoths migrateto mountainregions each summerto feed at nighton the nectarof alpineand subalpineflowers, and during the day they seek shelterunder various rock formations. Grizzly bears were observed feeding almost exclusively on moths up to 3 months each summerat the 10 moth-aggregationareas we identified. Fifty-one different grizzly bears were observed feeding at 4 of these areas during a single day in August 1991. Army cutworm moths are a preferred source of nutrition for many grizzly bears in the Yellowstone ecosystem and represent a high quality food that is available during hyperphagia.

Int. Conf. Bear Res. and Manage.9(1):389-399

In 1986 the InteragencyGrizzly Bear StudyTeam bear RecoveryZone for the Yellowstoneecosystem. (IGBST)first observed radio-instrumented grizzly bears Additionalsurvey work occurred adjacent to the excavatingand presumably feeding in alpinetalus fields primarystudy area, in the GallatinNational Forest to located in the southeastportion of the Yellowstone the northwest,in the northwestcomer of Yellowstone ecosystem. Subsequentvisits to these sites in 1987 NationalPark, and in the northernmountains of Grand revealedthat bears were feeding on aggregationsof TetonNational Park to the southwest(Fig. 1). armycutworm moths (Mattson et al. 1991). In 1988 The AbsarokaMountain Range originatedfrom we began a long-term project to investigatethe volcanicactivity during the LateEocene and Oligocene ecologicalrelationships between grizzly bears and army and variesin elevationfrom about2,000 to 4,000 m. cutwormmoths in the Yellowstoneecosystem. This The range is generally forested with deep valley papercovers some of the resultsof the first4 yearsof bottoms and numerous mountain peaks above this project. timberline. This projectwas fundedby the YellowstoneGrizzly The alpineterrain has been shapedby distinctglacial Foundationand the InteragencyGrizzly Bear Study erosion since the Pleistocene. Periglacialmesas are Team, with additional support from the Stanley distributedthroughout this regionand are subjectedto ResearchFoundation, the Cinnabar Foundation, and the strong winds, short growing seasons, inclement JamesF. SeedFoundation. Thanks are extended to B. weather,young soils, and wide temperaturegradients, Kondratiefffor moth identifications,B. Davitt for whichlimit the extentof plantcommunities. Extreme energeticanalysis, J. Claarfor assistancein the study topographicrelief occurs at the interfacebetween mesa design,and M. Menzel,D. Whittemore,D. Amato,D. landformsand parent rock, such as glacial cirque Rahbany, C. Mills, and S. Downing for their headwalls, cols, aretes, horns, and isolated rock fieldwork. precipices.This terrain is characterizedby andesiteand andesitic tuff rock formationsexposed chiefly by glaciation,mass wasting, and wind andstream erosion. STUDY AREA Talus fields located directly below parent rock The grizzlybear Recovery Zone in the Yellowstone formationscontain angular rocks of various size, ecosystem is approximately2,460,500 ha and beginningwith fine debrisat the uppermostreaches of encompassesYellowstone National Park and adjacent the steeper slopes and generally increasingin size portionsof Idaho,Montana, and Wyoming (Fig. 1). A fartherdownslope to largeboulders on the moregradual high plateau occupies the central portion and is slopesnear the bottomof the talusfields. Some talus surroundedby mountainswithin 5 differentnational is bare(< 10% plantcover) due to the activeprocess of forests. Climate, physiography, and vegetative masswasting, which precludes substantial soil andplant characteristicsof this area were describedby Knight development,and containsmall mudflow channels and and Eberhardt(1985). streamsof glacialand snow meltwaters. The primary study area was in the Absaroka Partiallyvegetated talus (10-50%plant cover) also Mountainslocated in the southeastportion of the grizzly occurson moderateslopes thatundergo less seasonal 390 Int. Conf.Bear Res. and Manage.9(1) 1994

( :IPrimary Study Area

AdditionalAreas Surveyed

Fig. 1. Study area in the Yellowstone ecosystem. movement providing greater stability for soil and this zone and extends to the valley bottoms below. vegetationdevelopment. Vegetative talus ( > 50 % plant cover) occurs within the areas of the least seasonal movement below parent rock, which allows for the METHODS greatest soil development. Alpine meadows containingthickly vegetatedareas of Bear Ecology gentle to moderateslope are common below the various Data were collected from ground and aerial talus formations. Moist and developed inseptisol soils, observationsof bears feeding at alpine moth-aggregation protection from strong winds, and relatively longer areas from June to November, 1988-91. Time, date, growing seasons (due to early retreat of annual and location were recorded for each observation along snowpack) typically yield taller and more diversified with species, age and sex class, and family group communities of vascular plants in alpine meadows composition. Aerial observations were made from compared to that found on periglacial mesas and fixed-wing aircraft as part of ongoing interagency partiallyvegetated and vegetated talus. observation flights and relocation flights for Krummholz occurs at the lower margins of alpine radio-instrumentedbears. The location of each bear meadows just above treeline. This zone contains observation was plotted on U.S. Geological Survey stuntedarborescent vascular plants that generallydo not (USGS) 7.5 minutetopographic maps. Bear feces were exceed 2-3 m in height. Closed forest begins just below collected from these areas for content analysis each fall GRIZZLYBEARS-ARMY CUTWORM MOTHS * French et al. 391 after the bears had left. RESULTS Observationswere made nonintrusivelyfrom distant vantage points using binocularsand spotting scopes. A Moth Ecology 16 mm motion picture cameraand a telephotolens were Behavior.--All moths collected at the various alpine used to film some bears for furtherbehavioral analysis. aggregation sites were identified as the army cutworm Some bears were identified as individuals by their moth (Euxoa auxiliaris). Their arrival each year distinctive physical features. correspondedto the onset of the flowering season for Subadults were distinguished by their smaller size the alpine flower communitiesnear the end of June and and shorterbody length comparedto adult females with the first part of July. They fed at night on the nectar of offspring observed in the same area. Adult males were several flowers, including green bluebells (Mertensia identifiedby genitaliaand were also easily distinguished lanceolata), thick-leavedgroundsel (Senecio crassulus), by their larger size relative to adultfemales with young. mules-ears (Wyethia amplexicaulis Nutt.), and In some cases, subadult males were identified by elephant's head (Pedicularis groenlandica). Moths genitalia. were rarely observed flying or feeding on windy or Seasonaluse, daily activitypatterns, activity budgets, rainy nights. and feeding behavior of bears feeding on moths were During the day, moths aggregatedunder the various studied at 2 areas during 1991. A 15-minutesampling rock formations, presumably to seek shelter from the period was used to determinethe activity budgets of all heat and frequent rain and/or hailstorms while age and sex classes of grizzly bears observed while metabolizing nectar. They were located in relatively feeding on aggregations of army cutworm moths. A large clusters distributednonuniformly throughout the 10-second scan intervalwas used to record the category talus and boulder fields. Moth-aggregationsites were of activity throughout the sampling period. Activity altered on a daily basis from extensive bear categories included digging for moths, walking between excavations. Some sites were abandoned by moths excavation sites, looking for moths, and eating moths. when the runoff from meltwatersceased but other sites became available when large perennial snowfields Moth Ecology receded. Moth activity was determinedby direct observations Moths were generally inactive while aggregated made during the day and at night. Some moths were under rocks although some appeared to be drinking captured and manually killed for species identification from glacial or snow meltwaters, and copulation was and biochemical analysis. The physical characteristics occasionally observed. This sexual behavior was first of moth-aggregation sites were recorded such as observed in mid-July and continued until geological formation, rock type, size and depth, slope, mid-September,and was also observed at night while aspect, elevation, temperaturegradient, and proximity moths fed. to sources of water and vegetation. Moth samples collected in late August were found to Moth samples were oven dried at 100?C for 24 hours contain approximately72% fat, 28% protein, and less prior to nutritiveanalysis. Crude proteinwas measured than 1% carbohydrate. The high fat content produced by the macro-Kjeldahl technique, crude fat was a relativelyhigh gross energy (7.9 Kcal/gm) for moths determined using chloroform-methanolextraction in a in late summer comparedto other available bear foods Gold fisch apparatus,and gross energy was determined (Table 1). in an adiabaticbomb calorimeter. By mid to late September each year, we rarely Moth distributions within talus fields were observed moths feeding at night and by the first of determinedby establishing 10 m x 10 m plots with a October we no longer found moths at the aggregation 2-m grid within each plot. The talus was excavated sites in the talus slopes. The disappearanceof moths approximatelythe width of a hand at each grid point from the alpine regions each fall corresponded to the until soil or ice were encountered and each moth was desiccation of alpine flowers or the onset of the first collected and counted. Thermal gradients were substantialsnowfall. determined by measuring air temperatures taken at Site Characteristics.-Sixty-eight moth-aggregation various depths in the talus. To avoid disturbingbears, sites were identified in the alpine regions of the study these collections and measurements were only made area, 38 by aerial observation, 14 by ground when no bears were observed on or adjacentto the talus observation, and 16 by a combinationof both methods. fields examined. We examined 25 of these sites from 1988 to 1991 and 392 Int. Conf.Bear Res. and Manage.9(1) 1994

moths Table 1. Gross energy of army cutworm compared with more time warming up before flight, and as a other bear foods as reported by Pritchard and Robbins (1990). consequence, usually fled by crawling deeper into the rock interstices. This mode of escape concentrated Fooditem Kcal/gm moths at the bottom of the excavations when they were Blueberries 4.47 preventedfrom going down fartherby soil or ice. White clover 4.83 Bear Ecology Ground squirrels 5.28 Bearswere observed feeding on mothsat 64 different Cutthroattrout 5.71 aggregationsites. These sites were distributedin clustersof 2-9 each locatedaround a common Pinyon pine nuts 6.48 sites, mountainpeak or massif. We commonlyobserved Deer 7.32 bearsfeeding from one site to anotherwithin a cluster Army cutworm moths 7.91 and also noted that radio-instrumentedbears used multiplesites withina single clusterduring the year. We thereforedefined each discrete cluster of feed sites recordedtheir physical characteristics. The total size of as a single moth-feeding area. Based upon this each site varied from approximately 1 to 16 ha as definition, we identified 10 moth-feeding areas within estimated from topographic maps. Site elevations the primary study area. The mean elevation of the 10 varied from 2,768 to 3,645 m (x = 3,313 m), while the mountainpeaks that containedthese areas was 3,378 m slope ranged from 10 to 56? (x = 39). Ten sites were (SD + 191 m). The distance between each moth on NE aspects, 7 on SE aspects, 4 on NW aspects, and feeding area averaged about 17 km and varied from 10 4 on SW aspects. to 30 km. Aggregation sites examined by ground were located Grizzly and black bears were observed 507 times in bare talus fields immediatelybelow parentrock, such feeding at the 10 moth-aggregationareas. Grizzly bears as glacier cirques (12), cols (6), or aretes (4). The 3 constituted92% (290) of the 314 ground observations remaining sites were located at the lower portions of and 97% (188) of the 193 aerial observations. There talus within boulder fields. Two-thirds of the sites were only 12 observations of radio-instrumented were considered to be active with ongoing annual rock moth-eatinggrizzly bears during this study, 11 by air of deposition from above, while the remainderwere within 4 differentbears, and once by ground of an unidentified partiallyvegetative talus. Most of these sites contained bear with an inactive radio collar. uniform angular rock varying from 6 to 12 cm in We developed a model to predict the locations of diameter, but some sites had larger, nonuniform rock other moth-aggregation areas based upon the that varied in size from 82 to 200 cm. examinationof the known moth-aggregationareas used Water was common to all moth-aggregationsites. by bears. This model included isolated mountain Ten sites were immediately adjacent to perennial massifs containing parent rock that created both bare snowfields and 5 were in bare talus on top of glaciers. and partiallyvegetated talus fields below. Parent rock The 10 remaining sites were moistened by streams or formations include glacial cirque headwalls, cols, rivulets of glacial or snow meltwaters. Moth sites aretes, horns, and isolated rock precipices. Talus fields below small perennial snowfields were abandonedby were either adjacentto perennial snowfields or on top late summer after the snow had melted and the of glaciers, and were located near alpine meadows underlying soils had dried. containinga variety of flowering vascular plants. A temperaturegradient was found within each moth- We tested this model by surveying 4 additionalareas aggregation site examined. On sunny days the air that were not known to have either moth-aggregations temperatureoccasionally exceeded 40?C at the rock or bear-feeding activity. We found moth-aggregations 3. surface, but generally dropped to 20?C at soil surface at all 4 areas and evidence of bear feeding activity at area but the typically within 20-30 cm, or dropped as low as One was adjacent to the primary study northwest freezing at sites overlying ice fields. others were more than 100 km away in the Temperaturegradient affected the moths' ability to and southwest portions of the Yellowstone ecosystem escape when they were exposed during excavations. (Fig. 1). Moths exposed near the rock surface had to flap their Feeding Behavior.-Bear use of army cutwormmoths wings several seconds before flying away, but those occurred during a 2-3 month period each year during exposed fartherdown where it was cooler had to spend this study. During the first 3 years this was determined GRIZZLYBEARS-ARMY CUTWORM MOTHS * French et al. 393

by aerialobservations of bearsfeeding at all 10 known 4 occasionsduring a 34-day period, once by aerial areasthroughout July and August, continuinginto the observationand 3 by radio-telemetryflights. first part of September. Based upon ground The thirdsubadult male monitored during 1991 was observations at 2 areas in 1991, the duration of this locatedat yet anothermoth-aggregation area 4 times feedingbehavior was slightlylonger, beginning in late duringa 23-dayperiod, 2 by radio-telemetryflights and Juneand continuinguntil late September,with a peak 2 by aerialobservation. He moved approximately15 of activityobserved in August(Fig. 2). km to the valleybelow but returnedto the same moth- The durationof this feedingbehavior by individual aggregationarea about 3 weeks later where he was bears was determinedby monitoringthe locationand locatedan additional3 times duringa 15-dayperiod, feedingactivities of 4 moth-feedinggrizzly bears who once by aerialobservation and twice by radio-telemetry were radio-instrumentedduring this study. During flights. Therefore, he was located and observed 1990a radio-instrumentedadult female grizzly bear was feedingat this moth-aggregationarea for a totalspan of locatedat a singlemoth-aggregation area 9 timesduring 38 days. an 80-day period, 8 by radio telemetryand once by Bears fed on moths in a crepuscularpattern and aerialobservation. individualbears generallyfed twice each day. This In 1991, 3 radio-instrumentedsubadult male grizzly feeding patternwas observedin 38 of 41 individual bears were monitoredat 3 separatemoth-aggregation bearsat 1 moth-aggregationarea during the monthsof areas. The firstwas locatedat a moth-aggregationarea July and August in 1991. During the daytimethey 8 times during a 38-day period, twice by restedin daybedsgenerally located nearby. radio-telemetryflights and 6 by aerialobservation. The Theybegan their morning feeding sessions sometime secondwas locatedat anothermoth-aggregation area on in the early morninghours because they were already

40 -

- 35 *Ground Obs. N = 176

AerialObs. N -193 30- ]

25- Number of Bears 20 - Observed 15 -

10-

5-

11 11 ..H1 ' ' 41...... I.. n. . lh II m,I II l, ? B . . I L 0 inL l. l . . __ l I July August September

Fig. 2. Seasonal use of alpine aggregations of army cutworm moths by grizzly bears based upon ground and aerial observations during 1991. 394 Int. Conf.Bear Res. and Manage.9(1) 1994 on the talus slopes feeding at first light each day. They Table 2. Content of grizzly bear scats collected at one moth- area in 1991, N = 284. continued to feed until they startedbedding for the day aggregation between 0900 and 1100 hours. ------% Frequency %Volume We observed 103 instances of bedding by 83.80 34.5 moth-eatingbears and examined 73 daybeds. Eighteen Euxoa 79.23 34.12 were located above the talus slopes on rocky ledges or in caves, 16 were below the talus slopes within the Ant 6.69 0.19 krummholzor edge of the treeline, 13 were within the Bot fly larvae 2.11 0.05 talus fields, 9 were at the base of large boulders, 9 Graminoids 85.21 50.77 were on the adjacentmesas or alpine meadows, and 8 Carex 61.97 21.75 were on of snowfields. of these beds were top Many Poa 52.46 18.04 used by different bears during the course of a season 8.45 2.79 and from year to year. Deschampsia Bears resumed feeding again between 1600 and 1900 Juncus 7.04 2.70 hours each day and continued to feed until we left the Grass 3.17 0.99 observation area each evening at dusk around 2030 Grass-sedge 2.11 1.68 hours. We observed bears continue to feed on moths Melica 1.76 0.70 until 2300 hours on a moonlit in but night August 1991, Bromus 1.06 0.02 observationsbeyond this time were precludedby cloud Forbs 16.20 4.06 cover. 10.92 2.23 We recorded the feeding behavior of 61 unmarked Trifolium grizzly bears for a total of 146 hours and found that Epilobium 2.11 0.50 95% of their time was directed towards moths. Taraxacum 1.76 0.27 Grazingaccounted for 5 % of their feeding activitiesand Lomatium 1.06 0.42 appeared to be directed, whereas feeding on small Forb 1.06 0.27 mammals, such as (Ochotona picas princeps), Mammals 7.04 0.35 golden-mantled ground squirrels (Spermophilus Ursus 5.98 0.06 lateralis), 13-lined ground squirrels (Spermophilus sp. tridecemlineatus),and marmots(Marmotaflaviventris), Marmot 0.70 0.26 appearedto be incidental, and accounted for less than Pika 0.35 0.02 1% of feeding activities. We found no significant Debris 70.07 9.83 differences in this feeding patternbetween the various Dirt 69.72 6.03 and sex classes. age Rocks 37.32 3.59 We analyzed 284 bear scats collected from these Wood 0.35 0.15 same areas after the bears left in the fall. Moths were Needles 0.70 0.06 present in 84% of scats and composed 34% of the scat volume, whereas graminoids appeared in 85% of the Birds (Unidentified) 0.35 0.00 scat and composed 51 % of the scat volume (Table 2). Pine nuts (Pinus 0.70 0.54 The results of the activity budgets measured for 26 albicaulis) individual grizzly bears while feeding on moths Shrubs (Vacciniumsp.) 0.70 0.08 concentrated and indicated that moths were easily Sporophytes 0.70 0.02 obtained. Bears meanderedback and forth on the talus (Lycopodium) fields and appeared to locate moth clusters by scent. They spent several minutesexcavating and eatingbefore in moving on to search for another moth cluster. aerial-observationtechniques were more efficient at Approximately65 % of their time was spent eating and surveying for the presence of bear-feeding activity only 22% excavating. Review of film records revealed the remote and widely distributed moth-aggregation that bears commonly continued to feed on moths while areas, they did not appearto be as accurateas ground of excavating. Therefore, somewhere between 65% and observations for counting the total number 87 % of their total activity was spent eating during their moth-eatingbears at a particulararea. On 4 occasions, and pursuit of moths. concurrentaerial and ground observationsoccurred Ground Versus Aerial Observations.-Although only 59% (27 of 46) of the bears observed from the GRIZZLYBEARS-ARMY CUTWORM MOTHS * French et al. 395 groundwere observedfrom the air. males with an aggressive posture, ears back, teeth We also recordeda significantdifference in the age showing,and roaring, until the malesmoved away, at and sex structureof the bearsobserved feeding at one whichtime the femalesresumed feeding. of the moth-aggregationareas during ground Adult males were not aggressive towards adult observationscompared to aerialobservations. Families: femalesor their cubs and on several occasionsthey subadults:lone adults ratios were 16:45:39 during appearedto avoidinteracting with females with cubs by groundobservations compared to 28:7:64 fromthe air. feedingaround them or movingout of theirway. Adult The 62 groundobservations were based upon 57 field- males were also nonaggressivetowards subadults, observationdays between 22 June 1991 and 15 although subadults generally moved around an September1991 and the 14 aerialobservations were approaching adult male thereby avoiding overt basedupon 5 flightsmade between 22 July 1991and 19 aggressive encounters. September1991. There were only 4 aggressive encounters observed Numberof Bears Feeding on Moths.-It appearedthat between adult males. In each case an adult male a significantnumber of grizzlybears fed on mothsfor displacedanother adult to claim a specificexcavation at least partof each summerduring this study. As an site, using a similarmethod to thatadapted by females example, we observed 51 differentgrizzly bears feeding withcubs to displacesubadults. at 4 of the 10 moth-feedingareas on a single morning We recognizedseveral individual bears at 2 of the in August 1991. However, we did not estimatethe ground observationareas and were able to follow their total number of bears feeding on moths because the daily activities during most of the season in 1991. We moth-aggregation areas were not surveyed occasionallyobserved other bears, including all age and simultaneouslyduring this study and most of the sex classes, thatwere unfamiliarto us. We classified moth-eatingbears observedand monitoredwere not thesebears as transientsbecause they only remainedat instrumented. an areafor a day or 2 beforedisappearing. SocialDynamics. -It wascommon for several grizzly Transientbears, especially subadultsand females bears to feed on moths at the same time in close with young, generallyfed near the edge of the talus proximity. The largest congregationof bears we fields thus avoiding close contact with the resident observedat one areawas 23. On severaloccasions we bears. They also spent a larger amountof time looking observedas manyas 9 bearsfeeding within an areaof at otherbears whereas the residentbears spent most of approximately50-m diameter.During feeding periods, theirtime looking for pocketsof moths,excavating, or they traversedan area from side to side and up and feeding,and rarely looked at each other. The reaction down, resultingin a continuousintermingling among all of transientadult males to the residentswere mixed. age and sex classes. Some intermingledwith the residentswhereas others Despite the relatively large number of bears that were morecautious and remained near the edges. routinely fed in close proximity, very little social We made12 observationsof moth-eatingblack bears tension was apparent. Although agonistic behavior was from the groundat one of the moth-aggregationareas generally rare, females with cubs appeared to be the and 7 aerial observationsof black bears at 3 other most aggressive of all cohorts. They occasionally areas. Blackbears also fed crepuscularlybut all age displaced subadults from specific excavation sites by and sex classesappeared cautious and fed on the edges eitherwalking rapidly or runningtowards the site. The of the talusfields when grizzlies were present. Grizzly displacedsubadult generally began searchingfor new bears were not aggressive towards black bears on the excavationsites within 20 m. Only twice did we talus fields, although black bears avoided overt observeda femalewith cubs chasea subadultfor more aggressive interactionby retreatingwhen a grizzly than50 m from an excavationsite. approached. Femaleswith cubs routinelyfed in the presenceof adult males. Cubs-of-the-year fed independentlyand OtherActivities occasionally ventured more than 50 m away from their Bears spent the majorityof their active periods at the mother. On several occasions, we observed other bears moth-aggregationareas feeding (> 85 %), but play was come betweena femaleand her cubs with no apparent a significant behavioral component for both subadults reaction from the bears involved. However, twice we (10%) and family groups (6%). Snowfields appeared observedfemales charge adult males who camebetween to stimulateplay in bothgroups, and even adultmales her and her cubs. Althoughno physicalcontact was on rareoccasions (< 1%). made in eithercase, the femalesremained next to the Snowfieldswere also used for a variety of other 396 Int. Conf.Bear Res. and Manage.9(1) 1994

purposes. Bearsappeared to preferthem as travellanes (Pruess1967). Eachfall adultfemales lay theireggs in comparedto adjacentrock. They also ate snow, the soil at lowerelevations throughout the GreatPlains. commonlylaid on snow for brief rest periodsduring Larvae undergo several instars during winter theirfeeding sessions, and occasionally dug daybedsin hibernationand emerge in early spring as the army snow and restedduring the middleof the day. Several cutworm.The cutworm feeds on leafyplants, including adult males were also observedto lie down upon a commercial crops such as wheat, barley, oats, snowfield and thrusttheir pelvis repeatedlyinto the sugarbeets,and alfalfa, and are sprayedwith pesticides snow. to reducecrop damage (Morrill and Borchers1986). Steep snowfieldsextending to the upperregions of Afterfeeding for severalweeks, the cutwormlarvae the talus fields were used to gain access to the rocky form cells undergroundwhere they enter the pupal ledges and caves that were used for daybeds. Bears stageand emerge as adultmoths around the firstpart of were cautiouswhen they descendedextremely steep summer. Shortlythereafter, they migratein massive sectionsof snow and backeddown while keepingat numbersto distant alpine habitatswhere they stay least 1 paw firmlyanchored into the snow. Whenthe duringthe summer(Pepper 1932, Koerwitzand Pruess snowfieldbecame less steep,they commonly turned and 1964, Kendall 1981). They have been reportedto descendedhead first, sometimessliding or running,but travelas far as 470 km (300 miles) from the Plains when they began going too fast or nearedthe bottom each summer(Pruess 1967). The exact distancesthey they spun aroundbackwards and stoppedby forcing travel from the plains may be even greater. As an theirfront claws deep into the snow. example,a markingexperiment with the orientalarmy Throughouteach summer,individual bears in all age cutwormmoth (Pseudaletia separata) of Chinarevealed and sex classes were observedfeeding at severalsites they migratedat least 1,400 km (870 miles) each withineach of our study areas, sometimeswithin the summerto the mountains,and traveledat a rate of same day. Steep ridges and rock walls separating about150 to 200 km (95 to 125 miles)each day during severalof these feed sites were negotiatedwith little migration(Baker 1991). difficultyand bearshad well establishedtravel routes Earlier researchersbelieved that army cutworm they used when going from one site to anotherwithin mothsaestivated during their summer stay in the alpine an area. regions (Cook 1927, Pepper 1932, Chapmanet al. Hazards.- Becauseof the extremetopographical relief 1955)but subsequent studies found evidence to suggest and high elevationscharacteristic of the moth feeding they remainedactive (Pruess 1967, Kendall 1981). areas,bears were exposedto significantenvironmental Metabolicstudies provide strong evidencethat army andphysical hazards. Lightningstorms were common cutwormmoths are active duringthe summerin the to the study area throughoutthe summersand bears alpine. Kendall(1981) reportedthat the abdominalfat wereexposed to nearbyground strikes. However,they contentof armycutworm moths was about30% when continuedfeeding throughout the most severeweather they first arrivein the mountainsof Coloradoin the conditions,including lightning, hail, and snow, and springand increased linearly throughout the summerto were never observedto stop feeding and leave the about 70% by the time they left the mountainsin slopesto seek shelterduring such conditions. September. Rocksdislodged from natural erosion, from mountain The armycutworm moths we observedduring this sheep(Ovis canadensis),and frombears represented a studywere obviously active during their alpine stay and significantrisk of injuryfor moth-eatingbears. They also appearedto play an importantrole in alpine were usuallyaware of the frequentrockfalls above and ecology. They fed on and pollinateda variety of most of the time were able to jump clear of danger. flowers. Based upon these observationscombined However,we occasionallysaw bearswho werestruck, withKendall's metabolicstudies, we believe army sometimesby rock from above and sometimesby the cutworm moths convert ingested flower nectar to rocksthat they dislodged. Most of theseinjuries were body fat. Army cutwormmoths were also observed minorbut we observedan adultfemale black bear who being preyed upon by black and grizzly bears, was crushedto deathwhen the largeboulder she was coyotes (Canis latrans), ravens (Corvus corax), diggingunderneath became dislodged and fell on top of Americanpipits (Anthus spinoletta), mountain bluebirds her. (Sialia curricoides),owls (speciesnot identified),and bats (probablyMyotis lucifugus). DISCUSSION Both sexes of army cutwormmoths returnto the Army cutwormmoths have 1 generationper year Plainsin the fall wherethe adultfemales lay theireggs GRIZZLYBEARS-ARMY CUTWORM MOTHS * French et al. 397

(Pruess 1967). Only rare observations of copulation is difficult. have been reportedby moths in the mountains(Kendall Grizzly bears have been reported to feed on army 1981) but Pruess (1967) found that females returningto cutworm moths in talus slopes below a single peak in the Plains in the fall had mature eggs and concluded the Mission Mountains(Chapman et al. 1955; Servheen they mated sometime prior to their arrival. Based upon 1983; Klaver et al. 1986) but less than 10 bears have our observations of moths copulating both at nectar been observed feeding at this site each year during the feeding sites and at daytime aggregation sites, we past decade (James Claar, biologist, U.S. Dep. of believe that a significant amount of reproductive Agric. [USDA] Forest Service, pers. commun.). behavior occurs during their alpine stay. The rise in fat KatherineKendall (biologist, U.S. Dep. of Inter. Nat. content (Kendall 1981) not only provides energetic Park Serv., pers. commun.) has observed grizzly bears substrate needed for fall migration, but may also be excavating for extended periods in several talus slopes important for sexual maturation and thus stimulate in Glacier National Park and believes they are feeding reproductivebehavior. on moths. And moths identified only as "belongingto The army cutworm moth is probablya native species the Family "were reportedlyfed on by bears in the temperate zones in North America. Although in higher elevation rock slides in the Scapegoat taxonomic classificationsof moths have changed during Wilderness, although it was unclear how this feeding the past 100 years, the NorthernArmy Worm Moth is behavior was documented(Craighead et al. 1982). at least an ecological equivalentif not the same species There are insufficienthistorical records to determine as the army cutworm moth described by Holland in how long grizzly bears in the Yellowstone ecosystem 1903 (1968). This moth has been reported in North have fed on alpine aggregations of army cutworm America for a least 200 years and was distributedfrom moths. Although the first scientific documentationof the Atlantic to the Rocky Mountains and from Canada this feeding behavior in this ecosystem occurred in to Texas. It was first noted to cause major crop 1987, locations of radio-instrumented bears and damage in great numbers as early as 1743 in the New incidental aerial observations of unmarked bears in England area, and in 1861 caused massive crop rocky alpine regions suggest that some bears may have damages in southern Illinois and Missouri (Holland been feeding on moths since at least 1981 (Mattson et 1968). al. 1991). The army cutworm moth was well documentedin the There is some evidence that bears may have fed on general area of the Yellowstone ecosystem prior to the moths in the Yellowstone ecosystem long before it establishmentof large-scaleagriculture development and became apparentto the scientific community. Local humansettlement. Previously classified as Chorizagrotis outfitters we interviewed recalled seeing bears feeding auxiliaris, this moth was present in significantnumbers in the alpine talus on "moths," "bugs," and "insects" in several western regions in the first part of this since at least the early 1950s. Some reported they century, including Montana (Cooley 1908), Colorado killed several bears at these sites prior to the cessation (Gillette 1904), and Canada(Strickland 1916). of legal sport hunting in the Yellowstone ecosystem in Humans currentlyplay a majorrole in the ecology of 1975. However, legal hunting of grizzly bears may the army cutworm moth but the net effect this has on have suppressedthis feeding behavior by reducing the the availabilityof alpine aggregationsof army cutworm concentrationof bears living near the known current moths for bears is not known. Extensive agriculture moth-feeding areas and opportunistically killing development of the Great Plains during the past 100 moth-feeding bears who were vulnerable to hunters years has dramaticallyaltered the native vegetation and when they fed on the open talus slopes. thus the army cutworms' lowland habitat. And current We are unable to determine if grizzly bear use of pest control measures directed toward the larvae may army cutworm moths in the Yellowstone ecosystem has influence the numbers of adults that migrate to the increased during the past few years. The effort to alpine regions each summer. observe this feeding behavior has obviously increased Grizzly bear use of alpine aggregations of army in recent years, both from the ground and from the air. cutworm moths is not confined to the Yellowstone Observationflights made prior to this study did not take ecosystem, but the magnitudeof this feeding behavior into consideration the daily feeding patterns. As an is greater than reportedin other places, in terms of the example, in 1991 more than 50% of the aerial total number of moth-aggregationareas and the number observations of bears at the moth-aggregation areas of moth-eating bears. However, other areas have not occurred between the hours of 0900 and 1100, the been studied as extensively so a direct comparison period when most bears became inactive. Therefore it 398 Int. Conf.Bear Res. and Manage.9(1) 1994

is not possible to accuratelycalculate and compare thangrasses. observationsuccess rates from previous years to This study clearly demonstrates that alpine quantifyany changein the observationof this feeding aggregationsof armycutworm moths are an important behavior. food for many grizzly bears in the Yellowstone Aerial monitoring of this feeding behavior is ecosystem. They also appearespecially important to definitely more efficient than ground observations adultfemales since each moth-feedingarea contained althoughaerial observers who pass over a slopewithin from 1 to 3 distinctivefamily groups throughout each a few minutesare not as accurateas groundobservers mothfeeding season during this study. Mothsare also who spend several hours scanningthe slopes with valuableto bears because they are high in energy, spottingscopes. Aerialmonitoring is possiblythe best abundant, concentrated,easily obtained, and are way to monitoradult female use of army cutworm availablein relativelyremote locations that serve to moths especially since monitoringappears biased reducepotential human-bear conflicts during the peak towardsthe observationof familygroups. tourist season. And finally, moths are important Mothaggregations appear to be importantfor all age because they are available during hyperphagia,a and sex classes of bears. Klaver et al. (1986) metabolicstage that begins in mid-Julywhen grizzly speculatedthat moth-aggregationareas in the Mission bears increasetheir feedingactivity to accumulatefat Mountainswere used primarilyby subordinatebears, reservesrequired for winterhibernation (Nelson et al. such as females with young and subadults,to avoid 1980). adultmales during the summermonths. Mattsonet al. (1991) speculatedthat subadults were underrepresented at the moth-aggregationareas in the Absaroka RECOMMENDATIONS Mountainsbased upon aerialobservations. However, we demonstratedthat aerial observations were Moth Ecology inadequateto accuratelydifferentiate between single Furtherstudy is needed to determinewhere army adultand single subadultbears. Therefore,our study cutwormmoths originate prior to migrationand how did not supporteither of thesehypotheses. the agriculturalpractices in these areas affect annual It was interestingto notethe discrepancybetween the mothpopulations. Of particularconcern is thepotential observed feeding activity and the scat content. thatadult moths may harbor pesticide residues that may Ninety-fivepercent of theobserved feeding activity was be contaminatingthe alpineecosystems and watersheds directedtowards moths, and although moths were found andmay also bioconcentratein individualgrizzly bears in 84% of the collected scat, they composed a who consumelarge numbers of mothseach summer. disproportionatelylow 34% of the scat volume. In It is also importantto developa methodto monitor contrastonly 5% of the observedfeeding activity was thelong-term availability of armycutworm moths in the directedtowards grazing, but graminoids were found in Yellowstoneecosystem and determinehow climate 85% of the scats and represented51% of the scat affects the annualabundance and thus availabilityof volume. mothsto bears. Mothtraps located in adjacentalpine There are several possible explanationsfor the and subalpineflower communitiesappears to be the discrepancybetween observed feeding activity and scat most practicaland reliablemethod. A grid systemof content. Observationsof feeding activity were traps could be placed in selected areas to ensure precludedat nightwhen bears may have movedoff the adequatesampling since the period of nectarproduction rockyslopes to feed on grasses. Bearsmay have also variesamong plant species and is determinedin partby fed more on grasseswhen they left the alpineregions slope, aspect,elevation, canopy, and moisture. Moth out of our sight but defecatedwhen they returned. collectionsshould also be madeat regularintervals each Moth scats may have been more likely to become year to determinean index of seasonal and annual buriedwithin the extensiveexcavations that occurred in abundance.This methodcould also be used to survey the talus fields. However, we believe the major other potentialareas to determinethe presence of explanation of this discrepancy lies within the moths. differences between the digestibility and caloric concentrationof the 2 foods. Moths are a more Bear Ecology concentratedform of energy and are probablymore A trappingand monitoringproject is needed to digestiblethan grasses and thereforecontribute more evaluate the population and social dynamics of ingested calories per unit weight and volume moth-feedinggrizzly bears, and to determinetheir GRIZZLYBEARS-ARMY CUTWORM MOTHS * French et al. 399

movements, distribution, and habitat use to further HOLLAND,W.J. 1968. The moth book. Dover Publications, define the significance of this feeding behavior. Inc., New York, N.Y. 479pp. A survey is needed to locate other moth-aggregation KENDALL,D.M. 1981. Bionomics of Euxoa auxiliaris Grote areas in this ecosystem that may be used by bears. (: octuidae) in the Rocky Mountains and Until this survey is completed, we feel that any area comparisons with two resident species of alpine moths. that meets the criteriaof our predictivemodel should be M.S. Thesis, Colo. State Univ., Ft. Collins. 88pp. excluded from developmentsor other activitiesthat may KLAVER,R.W., J.J. CLAAR,D.B. ROCKWELL,H.R. MAYS, preclude bear use until it has been fully evaluated. ANDC.F. ACEVEDO.1986. Grizzly bears, insects, and Observationflights are useful to monitor the level of people: bear management in the McDonald Peak region, bear use at known moth-aggregationareas. However, Montana. Pages 205-211 in G.P. Contreras and K.E. based upon the crepuscularfeeding patternof bears, we Evans, compilers. Proc. of the Grizzly Bear Habitat recommend that observation flights be made at first Symposium. USDA For. Serv. Rep. INT-207. light before bears become inactive and less visible KNIGHT,R.R., ANDEBERHARDT, L.L. 1985. Population during the day to increase aerial observationefficiency. dynamics of Yellowstone grizzly bears. Ecology Since transientbears, especially adult males, may use 66:323-334. more than 1 moth feeding area, we furtherrecommend KOWERWITZ,F.L., ANDK.P. PRUESS. 1964. Migratory that observation flights be made over all known areas potential of the army cutworm. J. Kan. Entomol. Soc. on the same day to preclude duplicate counts. This 37:234-239. would allow a more accuratedetermination of the total MATTSON,D.J., C.M. GILLIN,S.A. BENSON,AND R.R. use of these areas by the Yellowstone grizzly bear KNIGHT. 1991. Bear feeding activity at alpine population. aggregation sites in the Yellowstone ecosystem. Can. J. Zool. 69:2430-2435. MORRILL,W.L., ANDT. BORCHERS.1986. Importantpest LITERATURECITED insects in Montana-thearmy cutworm. Entomology Series BAKER,R.R. 1991. Butterflies and moths. Pages 46-55 in 9. Montana Agric. Exper. Sta., Mont. St. Univ., R.R. Baker, ed. Fantasticjourneys-the marvels of Bozeman, June 1986. migration. Welden Owen Inc., London, England. NELSON,R.A., G.E. FOLK,E.W. PFEIFFER,J.J. CRAIGHEAD, CHAPMAN,J.A., J.I. ROMER,ANDJ. STARK. 1955. Ladybird C.J. JONKEL,AND D.M. WELLIK. 1980. Behavioral, beetle and cutworm adults army as food for grizzly bears biochemistry, and hibernationin black, grizzly and polar in Montana. Ecology 36(1):156-158. bears. Int. Conf. Bear Res. and Manage. 5:284-290. J.S. ANDG.B. CRAIGHEAD,J.J., SUMNER, SCAGGS.1982. A PEPPER,J.H. 1932. Observationson a unidirectionalflight definitive for of system analysis grizzly bear habitat and of army cutworm moths and their possible bearing on other wilderness resources utilizing Landsat multispectral aestivation. The Can. Entomologist 64(11):241-242. and imagery computertechnology. Wildl.-WildlandsInst. PRITCHARD,G.T., ANDC.T. ROBBINS.1990. Digestive and No.1. Monogr. Appendix page 226. Univ. Montana, metabolic efficiencies of grizzly and black bears. Can. J. Missoula. Zool. 68:1645-1651. COOK,W.C. 1927. Studies on the of ecology Montana PRUESS,K.P. 1967. Migration of the army cutworm, cutworms (Phalaenidae):geographic distribution, seasonal Chorizagrotis auxiliaris (Lepidoptera:Noctuidae). I. succession and local distributionof moths as affected by Evidence for a migration. Annals of the Entomological cultivation in the "Nigger Hollow" area. Ecology Soc. of Amer. 60(5):910-920. 8(2):158-173. SERVHEEN,C.W. 1983. Grizzly bear food habits, R.A. COOLEY, 1908. Fifth annual report of the state movements, and habitat selection in the Mission of Montana. An entomologist army cutworm, Mountains,Montana. J. Wildl. Manage. 47(4): 1026-1035. Chorizagrotis auxiliaris. Mont. Stn. Agric. Exp. Bull. STRICKLAND,E.H. 1916. The army cutworm (Chorizagrotis 71:136-147. auxiliaris Grote). Can. Dep. Agric., Ent. Branch Bull. GILLETTE,C.P. 1904. Some of the more importantinsects. 13:31. Colo. Agric. Exp. Stn. Bull. 94:3-16.