DIETSOF ASIATICBLACK IN TAIWAN, WITH METHODOLOGICAL AND GEOGRAPHICALCOMPARISONS

MEI-HSIUHWANG,' Conservation Biology Program, University of Minnesota,1980 McNeal Hall, 1985 Buford Avenue, St. Paul,MN 55108,USA, email: [email protected] DAVIDL. GARSHELIS, Minnesota Department of NaturalResources, 1201 East Highway 2, GrandRapids, MN 55744, USA, email: [email protected] YINGWANG, Department of Biology,National Taiwan Normal University, 88, Section4, TingchouRoad, Taipei 116, Taiwan Republic of ,email: [email protected]

Abstract:Few studies have been conducted on thediet of bearsin thetropics. During 1998-2000, we studiedthe diet of Asiaticblack bears ( thibetanusformosanus)in Yushan National Park, Taiwan, on theTropic of Cancer.We used 3 methodsto investigatetheir diet: scats,feeding sign, andinterviews with indigenous hunters. Most scats (n = 654) werefound during autumn and early winter, when bears congregated in an areaof abundantoaks and foraged mainly on acorns. In 1999,when acorns were less abundant, bears preyed more frequently on ungulates.We found fewer scats(n = 37) duringsummer, when bears were more dispersed and foraged on soft mast,and only 2 scatsduring spring, both containing mainly greenvegetation. These findings were corroborated by bearfeeding sign that we found,including hundreds of oakand soft mast-producingtrees withbroken branches, some forming a structureresembling a nest. Wealso found evidence of bearsfeeding on carrionand insects. It was more difficultto determineseasonality of dietusing sign, although this technique yielded a largerlist of bearfoods (n = 26) thanscat analysis (n = 19). However,the largest tally of bearfoods (n = 70) was obtainedby interviewingindigenous people who hunted in or nearthe park. Data generated fromthese interviews were less quantitativebut covered a muchbroader area and time span and provided better information on the diversityand seasonalityof thediet than that obtained from scats and sign. The3 methods,each somewhat biased and deficient, yielded complimentary views of beardiets. Ourresults were consistent with other studies of Asiaticblack bears in broadleafforests; these bears were omnivorous, opportunistic, and heavilydependent on hardmast. In ourstudy, however, bears consumed medium-sized ungulates more frequently than in otherareas.

Ursus13:111-125 (2002)

Keywords: Asiatic black , broadleaf forest, diet, feeding behavior, feeding sign, oak mast, predation, scat analysis, traditional ecological knowl- edge,Ursus thibetanusformosanus

Whereasinnumerable studies of food habitshave been tify feeding sign. Thus, feces have been the cornerstone conductedon Americanblack bears (Ursus americanus), for informationon diets of black bears. much less data are available on the diet of its apparent Feces may be foundwith relativeease andoften in abun- ecological counterpartand close taxonomicrelative (Tal- dance, especially where bears exist at high density and bot and Shields 1996), the Asiatic blackbear. Therehave routinelyuse trails(Matthews 1977). This, however,was been several studies in Japan(Takada 1979, Nozaki et al. not the case in Taiwan. The density of bearsin Taiwanis 1983, Torii 1989, Naganawa and Koyama 1994, unknown, but sighting information suggests it is low Mizoguchi et al. 1996, Hashimoto and Takatsuki1997, (Wanget al. 1993). Moreover,for thatreason and possi- Horiuchi et al. 2000, Huygens and Hayashi 2001), plus bly othersdiscussed herein, scats arenot commonlyfound some from Russia (Bromlei 1973), (Schaller 1969, on trails. Therefore,we could not rely on scat analysis Manjrekar1989, Saberwal 1989), and mainland China alone to investigate the food habits of Formosanblack (Wu 1983, Wang 1988, Schaller et al. 1989, Chen 1991, bears. Consequently,we employed 2 additional meth- Reid et al. 1991, Ma et al. 1994), but these representa ods: quantificationof feeding sign and compilation of small portion of this species' total range. Moreover,re- traditionalecological knowledge from interviews with sults of many of these studiesare not availablein English. indigenoushunters. Here, we add new data on the diet of Formosan black Our objectives were to: (1) describe seasonal and an- a bears, subspeciesinhabiting Taiwan. The only previous nual diets of Formosan black bears, (2) compare tech- informationon the diet andforaging behavior of Formosan niquesfor assessing beardiets, and (3) compareresults of black bears was obtained from feeding naturalfoods to a dietary studies of black bears from differentgeographic captive individual (Wang et al. 1992, Hwang and Wang areas. 1993). Other studies of bear diets have relied mainly on scat analyses. Some studies obtainedancillary dietary infor- STUDYAREA mation from analyses of collected stomach contents, ex- Taiwan is a 36,000-km2island located off the eastern aminationof or feeding sign, even observationsof bears coast of China. The island is characterizedby a high di- observationsof feeding. However, forest-dwellingbears versityof faunaand flora, due to its locationon the Tropic are generallyrare, and therehas been little effort to quan- of Cancer (23050' N) and rugged mountainousterrain, which dominatesthe interior. 1 cite as:Hwang, M-H., D.L. Garshelis, and Y. Wang. YushanNational Park (YNP; 23019'N, 121?10'E), es- 112 Ursus 13:2002 tablished in 1985, comprises 1,055 km2 of the Central japonica association), and temperate mixed forest MountainRange and is the largest national park in Tai- (Chamaecyparisformosensis-Pasania kawakamiiasso- wan (Fig. 1). The elevation ranges from 300 m above sea ciation). Besides black bears, 12 species of medium-large level to the Yushan main peak at 3,952 m, the highest inhabited the park, the most prevalent being mountain in northeast Asia. Two-thirds of the park is muntjac (Muntiacus reevesi), Formosan serow above 2,000 m in elevation and there are >30 mountain (Naemorhedus swinhoei), Formosan wild boar (Sus peaks >3,000 m. Deep valleys and steep slopes (72% of scrofa), sambar deer (Cervus unicolor), and Formosan terrain>55?) cause numerous landslides and waterfalls macaque (Macaca cyclopis) (Kou 1999). (S-J. Chen 1989). No hunting or human settlementis al- Ourprimary study area was in southeasternYNP, within lowed within the park. an elevation range of 300-2,800 m, in the watershed of Y-F. Chen (1989) described 6 plant zones in YNP re- the LakulakuRiver. Except for routinetrail maintenance lated to elevation: broadleafforest (300-1,800 m, mainly and construction of bridges, this area had little human consisting of Lauraceae spp. and Fagaceae spp.), activity during our study because visitors were banned Chamaecyparis(1,800-2,500 m), Tsugachinensis (2,500- from most of the area. Crumbled stone walls, terraced 3,000 m), Abies kawakamii(3,000-3,500 m), subalpine slopes, small clearings, and overgrown trails existed as shrub(>3,500 m), and an alpine herbaceouszone (>3,800 relicts of previous occupants of the area, including ab- m). In southeasternYNP, Kou (1999) reported527 spe- original people and Japanese. Our focus during the au- cies of vascular plants belonging to 360 genera and 125 tumnand winterwas within a once-settledhillside known families. He classified the vegetation zones in the areaas as Daphan, where we conducted most of our bear trap- subtropicalevergreen broadleaf forest (Cyclobalanopsis ping. This area,a 3-day hike from the parkentrance (Fig. glauca-Machilus philippinenes association),warm-tem- 1), was rich in oak trees (primarily Cyclobalanopsis perate evergreen broadleaf forest (Cyclobalanopsis glauca), which attractedbears during autumn. glauca-Pheobeformosana association), temperateever- Averageannual rainfall during this studywas 2,710 mm green broadleaf forest (Pasania kawakamii-Machilus at 500 m elevation (lowest in Jan: 39 mm; highest in Oct: 816 mm). The monsoon rainy season occurred during May-October,with typhoonsmost prevalentduring July- October. Snow occurred at elevations >3,000 m during December-March. Monthly mean daily temperatures were warmestin July (23.8?C) and coldest in Januaryand February(13.8?C). Yearlyaverage temperature was 10?C at 2,500-3,000 m and 5?C at 3,500 m.

METHODS This study, part of a larger ecological study in which bears were captured and radiotracked, was conducted duringJuly 1998-December 2000. We used 3 approaches for ascertainingbear diets: (1) scat analysis, (2) observa- tions of feeding sign, and (3) interviews with indigenous people who lived near our study area and had an intimate association with the forest. During fieldwork we also observed and recorded annual and seasonal changes in availabilityof some fruits that were thoughtto be impor- tantbear foods. We subjectivelyrated fruit production as high, moderate,or low, based on the amount of fruit ob- served on trees and bushes.

Scat Analysis We collected scats duringAugust 1998-December2000 Fig. 1. Yushan National Park study area, Taiwan, showing in the courseof trapping,radiotracking, and hiking in YNP. the oak-rich area where Asiatic (Formosan) black bears We also intensively searchedfor scats in areas where we in autumn and 3 districts congregated (Daphan) located radiocollaredbears or found bear sign. We re- encompassing indigenous villages where we interviewed residents about bears. corded the location of scats and estimatedhow long they DIETS OF ASIATICBLACK BEARS * Hwang et al. 113 had been on the groundbased on color, moisturecontent, freshness of the sign, location, elevation, and vegetation and degradation. type. For most scats, we identifiedfood items by eye in situ, but also collected a small sample which was sun dried in Interviewsof IndigenousPeople the field. When we could not identify items in the field, During July 1998-May 2000, we conducted in-depth we collected the whole scat and broughtit to a laboratory interviewswith indigenouspeople living in villages neigh- for examinationunder a dissectingmicroscope. We iden- boringYNP to gatherinformation about their knowledge tified individualfood items to species or the lowest pos- of bear diets. These were mainly Bunun tribal peoples sible taxon. To aid in our identifications,we observed who historicallylived in mountainousregions where they foraging sign aroundscat sites, relied on consultingbota- practicedslash andbur agriculture,hunting, and gather- nists and indigenous field guides, and used a reference ing wild foods. Bears,locally called"Tumad " werekilled, collectionof plantspecimens collected from the studysite. althoughthey were seldom a primarytarget of theirhunts. We grouped food items into 5 general categories: hard All bear hunting is now illegal, but neverthelesscontin- mast, soft mast, , vegetation (leaves, stems, or ues (Wang 1999; Wang and Hwang 1999, 2000). roots), and insect. We did not quantifybear hairs or de- We selected3 Bunundistricts for interviews. Interviews bris such as soil, stones, and wood particles, which we were conductedin 6 villages from each of the Touyanand assumed were ingested incidentally. We also excluded Juoshi districts,located southwestand southeastof YNP, scats with human-relatedfoods such as bait or items pil- respectively(Fig. 1). Althoughonly one village, Meishan fered from our researchstation. of Touyan,was located within the boundaryof YNP, the During 1998, we recordedonly frequencyof occurrence traditionalhunting territory of these villagerscovered most of differentfood items: FOi (%) = (n/N) x 100, where N (>75%) of the park. In the thirddistrict, Hidwan, which was the total numberof fecal samples and nithe number was outside the southernborder of YNP, we conducted of samplescontaining food item i. During 1999 and 2000, interviews in only 1 village, Wulu. we also estimatedthe relativevolume of each item in each The senior authorconducted all interviewsexcept one. scat, and averagedthese values among scats: RVi (%) = The interviewerspent time living in the villages, which V./ N, whereNwas the total numberof scat samplesand promotedmore congenial and frankdiscussions. Several ]V/ was the sum of the volumetricpercents for food item indigenousemployees of YNP helped us identify experi- i among all scats. enced hunterswho could provideinformation about bears. We separatedthe databoth seasonally and yearly. Sea- We also used a snowballsampling method (Bieracki and sons were determinedby changes in climate and plant Waldorf 1981) of referralsfrom one interviewee to an- phenology and were defined as follows: spring (Mar- other. All intervieweeswere men because only men were May), summer(Jun-Aug), autumn(Sep-Nov), and win- involved in traditionalhunting. ter (Dec-Jan) (no scats were collected in February). We Interviews were done face-to-face with open-ended latercombined autumn and winterand called this the oak questions. We asked aboutfoods thatbears eat, how they season,the periodwhen bears consumed mainly hard mast. obtain these foods, and how the person acquiredthis in- formation. In some cases, indigenouspark employees or Observationsof FeedingSign other indigenousinformants helped translateif the inter- While foragingin trees, Asiatic black bears sometimes viewee preferredusing their Bunun language. Some in- breakbranches to reachfruit on twigs. This behaviormay terviews were tape recordedto laterconfirm the accuracy lead to the formationof so-called tree platformsor nests, of translations. Many Bunun names for plants were ini- or more commonly,just obvious damage to the canopy. tially unfamiliarto us. Thus, duringinterviews we asked We recorded tree and shrub species with broken limbs for detaileddescriptions of such plants,their fruiting phe- caused by bearsor with bearclaw marksleading up to the nology, distribution,and habitat. We then confirmedthe canopy even if no brancheswere broken. Otherfeeding plants while working with villagers in the field, or we sign included remainsof ungulatesand beehives that we collected plant specimens and verified them with villag- identifiedas having been consumedby a bearfrom tracks ers later. These were eventually identified using field or other sign in the vicinity. We could not identify bear guides and assistancefrom botanists. feeding sign on herbaceous species because many me- diumand largemammals consumed these. We foundbear sign mainly duringthe course of other fieldwork. How- RESULTS ever,while radiotrackingbears, we sometimesspecifically searchedfor theirsign afterthey left an area. For all bear Scat Analysis feeding sign encountered, we recorded an estimate of We analyzed 693 bear scats: 268 in 1998, 95 in 1999, 114 Ursus 13:2002 and 330 in 2000. We excluded 25 scats containing hu- mainsand 1.2%contained insects. In thevolumetric analy- man-relatedfood. We found only 2 scats during spring sis, plant and food composed 93.0% and 7.0% of and 37 during summer;the remainder(654) were from scats, respectively. We identified 19 differentfood items: the oak season. Scats were found at 1,000-2,100 m el- 3 species of ungulates, 3 species of insects, and 13 spe- evation, although most (>90%) were at 1,100-1,600 m. cies of hard and soft fruits (Table 1). We did not try to Individualscats contained 1-4 food items, but 77% had identifyvegetation, such as grass,leaves, stems, androots, only 1 item and 19%had 2 items (mean = 1.3, SD = 0.5). which occurredin 2.7% of scats. Plant material (hard mast, soft mast, and other plant The 2 scats found during spring consisted mainly of parts) occurredin 99.3% of bear scats. Animal matter herbaceousplants (>95% relativevolume). One of these occurredin 12.3%of scats; 11.4%contained ungulate re- also containedinsects, and 1 containedmuntjac remains.

Table 1. Diet of Asiatic black bears based on scat analysis (n = 693 scats), observations of feeding sign (n > 600 climbed fruit trees, eaten carcasses, etc.), and Interviews (n = 70) with indigenous hunters (n > 440 reported items), Yushan National Park, Taiwan, 1998-2000. Methods' Consumed Category Common name Species name Scats Feeding sign Interviews parb Seasonc Hardmast Ring-cuppedoak Cyclobalanopsisglauca 534 >300 44 f a, w Devil tan oak Lithocarpuscastanopsisifolius 4 42 f a, w Smoothleaftan oak Pasania glabra 24 f a, w Long-leaf chinkapin Castanopsiscarlesii 20 f a, w Arishanoak Quercusstenophylloides 86 >200 13 f a, w Mori oak Quercusmorii 12 f a, w Tailuko oak Quercus tatakaensis 1 2 6 f a, w Taiwan walnut Juglans cathayensis 33 1 6 f a, w Long gland oak Cyclobalanopsislonginux 20-30 4 f a, w unknownoaks 2 Soft mast Nanmu Machilus spp. 30 50-100 35 f su Taiwan loquat Eriobotryadeflexa 9 16 16 f su, a Formosanapple Malusformosana 5 20 11 f a, w Indigenouscinnamon Cinnamomumosmophloeum 18 11 f su Giant taro Alocasia macrorrhizos 10 f su Taidon persimmon Diospyros oldhamii 2 6 f a, w Passion fruit Passiflora edulis 6 f su, a Taiwan cherry Prunus campanulata 6 4 f,fl sp Formosankiwi Actinidia callosa 4 f su, a Luzon viburnum Viburnumluzonicum 6 3 3 f a Oldhampersimmon Diospyros sasakii 2 f su, a Litsea Litsea spp. 3 2 f su Formosansugarplum Arengapinnata 2 f,s Shell ginger Alpinia speciosa 1 f a Taiwan phoebe Phoebeformosana 1 1 f su, Thunbergelaeagnus Elaeagnus thunbergii 1 f Autumnample tree Bischofia trifoliata f 1 f Wild mango Mangifera indica 1 su Cuming'swintergreen Gaultheriacumingiana 1 f f a Raspberry Rubusspp. sp, su, Rose Rosa spp. f Konishi neolitsea Neolitsea konishii f f Jelly-fig Ficus pumila Taiwan mulberry Morus acidosa 1 f f a Blueberry Vacciniumdonianum 2 su, Soap tree Sapindusmukorossi ~~~~~11 f su Rough-leavedtree Aphanantheaspera 1 f a Fetid securinega Securinega virosa 1 Mountainviburnum Viburnumpropinqum su Pouteria Planchonella obovata unknown 20 f 15 s sp, su Vegetation Dwarf bamboo Yushanianiitakayamensis Miscanthus 42 1 Japanesesilver-grass floridulus 1 Formosantaro Colocasiaformosana 15 , f, fl Taiwan kudzubean Pueraria lobata s Sword fern Nephrolepisauriculata 1 r Orchid Orchidaceae 1 1, s DIETSOF ASLATIC BLACK BEARS * Hwang et al. 115

Table 1. (continued) Methodsa Consumed Category Common name Species name Scats Feeding sign Interviews partb Seasonc Vegetation Taiwan dwarf banana Musaformosana 1 1 Fern Pteridophytaspp. 1 Nest fern Aspleniumnidus 1 1 Point-leavedmaple Acer insulare 5 Formosanalder Alnusformosana 4 Taiwan cedar Taiwaniacryptomerioides 2 su, sp Indiacharcialtrema Tremaorientalis 1 Taiwan red maple Acer morrisonense 1 unknown 7 7 Crop Peach 5 f Plum 5 f Prune 1 3 f su Pear 3 f Papaya 3 ff Bamboo shoot 3 s Corn 2 f Banana 2 f Italianmillet 2 f Sweet potato 1 r Insect Common bee Apidae 2 26 Hornet Vespidae 5 2 sp, a, w larva Coleoptera 4 Plant ball 1 Ant Hymenoptera 1 1 su Hymenoptera 1 Beetle Coleoptera 2 a, w Mammal Formosanmuntjac Muntiacusreevesi 794 2 common Formosanserow Naemorhedusswinhoei 79I 1 common Formosanwild boar Sus scrofa 1 occasional Sambardeer Cervus unicolor occasional Chinese civet Viverriculaindica I Bird Bird 1 Chicken 1 Fish Fish 12 Reptile Snake 2 Others Streamcrab 7 Streamshrimp 1 Earthworm 1 Mushroom 1 Human-related food 11 common sp, su, a, w a Each scat was a sample unit (values representthe numberof scats containingthe food item). The sample unit for sign was individualtrees, carcasses, beehives, etc. The sample unit for interviews was the mentionof a food item (values on table representthe numberof times each item was mentioned). b f = fruit, fl = flower, s = stem, 1 = leaf, r = root c sp = spring, su = summer,a = autumn,w = winter d Muntjacand serow hairs were not differentiatedin scats (79 scats contained 1 or both species).

During summer,scats were comprisedprimarily of fruits wild plums (Prunusspp.). The next most frequentlyob- (Fig. 2). Fruitsof nanmu(Machilus spp.) were mostpreva- served summer food category was mammals, followed lent (FO = 81%, RV = 77%) but these were found only in by vegetation (Fig. 2). No hardmast was found in scats 2000, when they were noticeably more abundantin the duringspring and summer. forest thanin the previous2 years. We also foundTaiwan During the oak seasons of all 3 years, bears focused loquats (Eriobotryadeflexa), Luzon viburnums(Vibur- intensively on hardmast (Fig. 3). We identified3 species num luzonicum), mountain viburnums (Viburnum of oaks consumedby bears (based mainly on the proxim- propinqum),rough-leaved trees (Aphananthe aspera), and ity of scats to a clump of oak trees): ring-cuppedoak 116 Ursus 13:2002

100 1.0

80 0.8

r 60 0.6 E 6o 0D OSoft mast >, >

40 - E Hard mast "1~~~~~~~~ ~~~~~~~~~a cr E]Mammal LL 20 0.2 l Vegetation

0 u lM 1- - Insect ,0 . i0.0 Summer Oak season Summer Oak season

Fig. 2. Seasonal changes (summer: Jun-Aug; oak season: Sep-Jan) in the diet of Asiatic black bears based on frequency of occurrence and relative volume of items found in scats (n = 693), Yushan National Park, Taiwan, 1998-2000.

100 1.0

80 0.8

60 - 0.6 E IEl Soft mast > ?' mast 40 0.4 >--I 04BUHard g' I: |oc Mammal 20 - 0.2 D Vegetation U Insect 0 I -- t - 0.0 1998 1999 2000 1999 2000

Fig. 3. Yearly variation in diets of Asiatic black bears based on frequency of occurrence and relative volume of items found in scats during the oak season (Sep-Jan) in Yushan National Park, Taiwan, 1998-2000. Only frequency of occurrence was measured in 1998.

(Cyclobalanopsis glauca), Arishan oak (Quercus tuations in productivityby these species. During 1998, stenophylloides),and Tailukooak (Q. tatakaensis). Wal- 1999, and2000, we ratedthe productionof acornsby ring- nuts (Juglans cathayensis) also were observed in scats, cuppedoak as high, low, and moderate,respectively; ac- althoughwe did not recognizethem until they were preva- cordingly,this species was most common in scats during lent in 1999. Bears began consuming hardmast in mid- 1998, followed by 2000, andabsent in 1999. Arishanoak Octoberand finished by mid-January,at which time they acornswere ratedas low-moderate,high, andlow in these left the Daphanoak feeding area. Mammals,principally same 3 years, and were identified in scats only in 1999. muntjacsand serows, were the second most commonfood However,Arishan oak is less commonin the Daphenarea item in scats during the oak season (FO = 10.9%, RV = thanring-cupped oak, so during1999 the high production 6.7%;Fig. 3). Soft fruits, vegetation, and insects each of Arishanoak did not make up for the low productionof occurredin <3% of scats and comprised<3% of scat vol- ring-cuppedoak; hence, in that year bears fed more on ume duringthe oak season. walnuts and mammals(Table 2; Fig. 3). The species andamount of hardmast consumed by bears variedby year (Table2), and this matchedthe yearlyfluc- FeedingSign DIETS OF ASIATICBLACK BEARS * Hwang et al. 117

Table 2. Frequency of occurrence (FO) and relative volume being damaged. One oak tree was used on at least 6 dif- (RV) of hard mast food items identified in scats of Asiatic ferent we do not know if these visits black bears oak seasons Yushan National occasions, although during (Sep-Jan), were the same bear. Under trees with abundant Park, Taiwan, 1998-2000. Species of oak mast were by fruit, identified mainly by trees in the vicinity of the scat. we foundup to 10 scats andmany torn off branches,often 1998 1999 2000 with numerousuneaten fruits . (n = 263) (n = 95) (n = 296) Non-vegetative items eaten by bears, as evidenced by Item FO FO RV FO RV their sign, includedtermites in rottenwood (n = 1), hon- All hardmast 97.3 94.7 75.9 98.3 95.2 eycombs in caves andunderground (n = 2), and carcasses Ringed-cupoak 97.3 0 0 94.3 94.2 of mammals, including 2 muntjacsand 1 serow, plus a Arishanoak 0 90.5 61.9 0 0 and wild boar that had been in hunters' Tailukooak 0 1.1 0 0 0 macaque caught Walnut Few' 28.4 14.0 1.6 1.0 snares. We never observed bears pursuingungulates, as a Walnutswere not identified in scats in 1998. reportedby some hunters. They may have scavenged al- readydead mammalsbecause carcasses(unrelated to hu- We found >600 occurrences of bear feeding sign on man activities) were common;we found 28, including 18 treesor shrubs,including 6 species thatproduce hard mast, muntjacs,9 serows, and 1 young macaque. 11 species thatproduce soft mast, 5 species of vegetation, Bears also sometimes took food from our research and 4 unknown species (Table 1). Commensuratewith camps, generallywhen we were away. They usually car- our resultsfrom scat analysis, the most prevalentfeeding ried this food to more secretivesites (e.g., undertall grass sign was associated with ring-cuppedand Arishan oaks or behindbig rocks or bushes). Theirsign includedscats, (>300 and >200 trees, respectively). Our observations vomit, scrapsof food and wrappers,and trampledresting suggestedthat bears preferred some oak species to others. sites. For example, we never found evidence of bearsclimbing Resting sites of bears (n = 46) often were associated large-leafoaks (Pasania ternaticupula),even in good mast with good food suppliesor some form of protection,such productionyears. Similarly,when both ring-cuppedand as tall grass,tree roots, stone caves, or rockoutcrops. Most long glandoaks (Cyclobalanopsislonginux) were produc- werejust depressionsin the vegetation. However,in some tive in the same generalarea in 2000, bearsused the former cases (n = 18) bears constructeda bowl-shaped nest by more frequently,based on numbersof trees with feeding intricately bending and twisting grasses (Miscanthus sign (Table 1). floridulus) or twigs (Rhododendronrubropilosum)(Fig. We frequentlyfound sign of feeding in nanmu trees, 4). The averagesize of these groundnests was about 110 particularlyduring the summerof 2000 when fruit pro- cm in outside diameter,60 cm inside diameter,and 30 cm ductionwas especially good. We could not always iden- in depth (n = 15 measured). tify the species of nanmu used by bears because field identificationof the genusMachilus is quitedifficult; how- ever, of the trees we could identify, most bear sign oc- curredon narrow-leavednanmus (M. japonica) and red nanmus(M. thunbergii).For some plantspecies, we could not identify which partsbears consumed or even whether bearsclimbed the treesfor foraging. These includedpoint- / leaved maple (Acer insulare), Formosan alder (Alnus formosana), Taiwan cedar (Taiwania cryptomerioides), and Indiacharcialtrema (Trema orientalis). 7-45 cm Numerousbroken branches in the tops of trees was evi- 41 dence of arborealfeeding behavior. Although we only L l.1 K -IA observedbears feeding in treestwice, we ascertainedhow 80-145 cm they fed from their sign. They often pulled fruit-bearing branchestoward them with their forepaws or broke big- ger limbs (e.g., >10-cm diameter)using theirteeth. Some- Fig. 4. Asiatic black bears in Yushan National Park, Taiwan, made nests blades times, in oak and nanmu trees, these broken branches bowl-shaped ground by twisting together of tall grass. These nests were not only a comfortable formed what looked like a platform or nest, especially resting spot, but they provided protective cover, which bears when the accumulationsof brokenbranches were stuffed sought especially when feeding near human-related food into a fork and trampledby the bear. We occasionally sources. These nests also may have served as a hiding observedseveral nest structuresin a tree. place from which they could ambush ungulate prey. single Repeated Measurements represent the range observed during this use by bearsof some treesresulted in >80%of the canopy study (n = 15 of 18 were measured). Drawing by W-J. Yu. 118 Ursus 13:2002

Interviewswith Indigenous People (Lithocarpuscastanopsisifolius) and long-leaf chinkapin We conducted70 interviews with indigenous hunters, (Castanopsiscarlesii), were more frequentlyreported by 27 from Juoshi,38 from Touyan,and 5 from Hidwandis- huntersin Touyan,who hunted at a generally higher el- trict. Most of the interviewees (96%) were Bununs who evation than those from Joushi. Conversely,low eleva- historically lived in the mountainousregions of central tion (<1,500 m) species, such as Taiwan loquat, were Taiwanand were skilled in hunting. The average age of reportedmore frequently by Juoshihunters. Nanmus, oc- respondentswas 57 (range= 30-80) years; 89% had ex- curringin a broadelevational range, were reportedas the periencehunting bears with guns or capturingthem with most importantsummer soft mast in both districts. wire snaresor traps. Most interviewees believed that bears were opportu- nistic foragersthat ate anythingedible, including plants, DISCUSSION live , carcasses, and human-relatedfoods. The most commonfood items firstmentioned by interviewees Comparisonswithin and across Species were wild ungulatesor theircarcasses from hunter's traps. Most bears are omnivorous, opportunistic feeders, Many huntersalso experienced bears taking food from whose diets can vary seasonally,yearly, geographically, habitat. These with differencesin their hunting huts. They believed bears had very short andby variables,along intestinesand small stomachs,which caused them to vomit investigative methodology, confound attemptsto make and when they ate too much. generalizations comparisons. With these caveats in we diets of Asi- Aside from ungulates and human-relatedfoods, indi- mind, compared vidual intervieweesmentioned 1-24 differentfood items atic blackbears inhabiting broadleaf forests similarto our was to view our results in the (mean = 6.3, SD = 4.4; 43 of 70 listed 6 or less); 70 inter- study site. Our aim here context of variationfound different views generateda list of 440 foods (including repeated among populations We also our with items). Plant foods included 9 species of hard mast, 24 of this species. compared findings Americanblack bear diets in similarforest habitat soft mast, 9 non-fruitingplants, and 10 crops (Table 1), types. Data on American black bears are often used as surro- plus 26 items (not all different)that we could not identify for Asiatic black bears for and habitat from the descriptionsor common names that the people gates population Miura used. Intervieweesmentioned hard mast most frequently assessments (Wanget al. 1994, Horino and 2000, Park because the 2 are considered to be (45% of listed items), and reportedthat bears intensively 2001) species data are often used oaks in autumnand winter. Soft mast, they said, genetically and ecologically similar, but for the latter. Our between the 2 was more diverse and mainly used duringthe summer. lacking comparisons are thus meant to at least on the level of Interviewees considered muntjacand serow the most species discern, how similar are. common ungulateprey of bearsand mentionedwild boar diet, ecologically they Across the of the Asiatic black bear, mate- andsambar deer less frequently.They suggestedthat bears range plant rial >80% of the overall diet and animals sometimespursued ungulates and preferred young or weak comprised formeda small of the diet;however, the individuals. Additionally,bears purportedlyambushed portion frequency of animalfood in our was than that in other ungulatesby hiding nearcliffs or in groundnests or even study higher areas 1979, Nozaki et al. 1983, Manjrekar1989, jumping on them as they passed undera tree. Bears also (Takada Schalleret al. 1989,Torii 1989, Reid et al. 1991,Naganawa were reportedto regularly use carrion, especially kills and 1994, et al. 1996, Hashimotoand made by yellow-throated martens (Martes flavigula), Koyama Mizoguchi Takatsuki1997, Horiuchiet al. 2000). The overall rela- which were reportedto consume mainly the innerparts. tive use of and animal (-10%) foods ap- Otherreported animal foods included 5 species of in- plant (-90%) to be similarfor Asiatic and Americanblack bears sects as well as fish, crabs, shrimp,civets, birds, chick- pears in broadleafforests Mattson[1998]). ens, snakes, earthworms,and mushrooms. Honey and (summarizedby Seasonal shifts follow the same generalpattern beehives frequentlywere indicated as a preferredfood. dietary for both Asiatic Lee et al. 1991, K. Lee et al. 1991, Interviewees reportedseeing evidence of bears digging (H. Reid et al. Hazumi 1994, Ma et al. 1994, Hashimoto out hives of commonbees (Apidae)and hornets (Vespidae) 1991, and Takatsuki1997, Horiuchiet al. 2000) and American from trees or undergroundand mentionedthat bears usu- black bears McDonald and Fuller [1993]). attackedthese nests at dawn, when stinging insects (reviewed by ally of black bears on were less active. In spring, both species depend mainly succulent In summer,soft fruits be- Huntersfrom the Juoshi and Touyandistricts showed green, vegetation. come the main In autumn,bears turn somewhat differentresponses, commensuratewith their dietarycomponent. to hardmast, if available,and both black bear species prefer respective hunting areas. Predominatemast species at certain of oak (Garshelis and Pelton 1981, higher elevations (>1,500 m), such as Devil tan oaks types DIETSOF ASIATICBLACK BEARS * Hwang et al. 119

Hashimoto and Takatsuki 1997, Huygens and Hayashi We suggest that in their high-gradingforaging strategy, 2001, this study). Hardmast, mainly of the Fagaceaefam- which may be necessaryto sustaina frugivoreof this size ily, composed 50-90% of autumndiets of black bears in (Welch et al. 1997), bears sometimes found it unprofit- broadleafforests throughoutAsia (Table3). In some ar- able to rummagethrough the fallen brancheswhen trees eas, fallen hardmast from the previous autumnwas con- with abundantmast were in the vicinity. sumedby bothAsiatic andAmerican black bears in spring When key foods become scarce, both Asiatic and and even summer (Nozaki et al. 1983; Tement 1995; American black bears turn to less preferred,less acces- Huygens et al., 2000, Diet and ecological aspects of Asi- sible, or riskier (e.g., human-related)foods. When hard atic black bear in the NorthernJapanese Alps, Shinshu mast crops fail, they tend to rely more on green vegeta- University,Matsumoto, Japan), although this did not oc- tion, soft mast, or agricultural crops (Hazumi and cur in our study because acors, without snow cover in Maruyama1986, 1987; Eiler et al. 1989; McDonaldet al. winter, either rotted or were consumed rapidly by other 1994; Kasbohm et al. 1995; Hashimota and Takatsuki animals. 1997; Vaughan2002). In Japan,black bears resortedto Some forest types producelittle hardmast. American tree cambium when naturalbroadleaf forests were con- black bears inhabitingsuch areas rely more on soft mast verted to plantationswith little available mast (Azuma duringautumn (Mattson 1998, Vaughan2002). No stud- and Torii 1980, Furubayashiet al. 1980, Hazumi 1994). ies of diet of Asiatic black bears have been conductedin Bears in our study faced with a poor acorn crop (1999) places lacking hard mast, which may explain why this increasedconsumption of walnutsand ungulates. species appears(but may not necessarilybe) the mosthard- The digestive system of bearsis betteradapted for meat mastdependent of all the bears(Mattson 1998). Ourstudy than more structurally-complexplant material(Bunnell confirmedthe importanceof hardmast in bear diets in a and Hamilton 1983, Pritchardand Robbins 1990, Hewitt forest where it is seasonally abundant.Hard mast is high andRobbins 1996). However,mammals compose a small in fat et al. (Landers 1979, Eagle and Pelton 1983, portionof blackbear diets, and insects seem an especially Kasbohm et al. 1995, Hashimoto and Takatsuki 1997, variablecomponent of the diet (Landerset al. 1979,Takada Kirkpatrickand Pekins 2002), and duringautumn the di- 1979, Beeman and Pelton 1980, Graberand White 1983, gestive system of bears becomes more able to assimilate Grenfell and Brody 1983, Torii 1989, Raine and Kansas fat (Brody and Pelton 1988). Bears also undergoa period 1990, Holcroft and Herrero 1991, Hashimoto and of pre-denninghyperphagia (Nelson et al. 1983) in which Takatsuki1997, Khramtsov1997). In our study, insects they substantiallyincrease their intake of food. In combi- were a very smallportion of the diet,but mammals seemed nation, these factorsresult in rapidweight gain, an adap- generallymore importantthan in other studies of Asiatic tation for hibernation. Our observations suggest that or Americanblack bears (Table3; Mattson 1998). Simi- Asiatic black bears in Taiwan behaved similarly, even lar to our study,Wu (1983) found thatAsiatic black bears these bearsdid not though hibernate(Hwang et al. 2000). in China preyed on ungulates when plant food was less One enigma is our observation of many fruit-laden available. Most studies of American black bears indi- branchesbeneath foragingtrees, apparentlyneglected by cated thatthey preyedmainly on young or weak prey ani- bears. Schaller and (1969) Bromlei (1973) both observed mals (Wilton 1983, Mathews and Porter 1988, Schwartz Asiatic black bearsdescending from trees to eat fruitsoff and Franzmann1991, Ballard 1992, Kunkel and Mech branches that either fell or were intentionally dropped. 1994, DeBruyn1997). We suggestthat the relativelysmall

Table 3. Autumn diets, based on scat of Asiatic black analysis, bears in broadleaf forests of different geographic regionsa, ordered from south to north. Bears in Taiwan (this study) were more reliant on vertebrate (ungulate) prey than in other areas. Diet composition (%) Sample Site Location Methodb size Vegetation Soft mast Hardmast Vertebrate Invertebrate Taiwan 121?E 23?N, FO/RV 654 1.8/1.6 2/0.8 85.8/90.5 9.6/6.7 0.8/0.4 China 103-105?E 31-33?N, RV 46 5 5 90 Few 0 India 75?E 34?N, RV 45 0 17.2 80.4 0 0.7 Gifu, 137?E Japan 36?N, RV 80 <0.1 14.2 84.1 <0.1 0.2 Nagano, 138?E Japan 37?N, RV 86 <0.1 34.9 63.1 0.1 1.8 Russia 42-50?N, 130-140?E FO 70 10.5 33 52.1 0.4 4 a Valuesobtained from studies in which >40 only published scats were collected over a period of >1 month duringautumn-early winter. Where data were groupedby weeks, months, or we recalculated years, an overall averagefor this season, weighting time periods by sample size. Sources - China: Schalleret al. India: (1989); Manjrekar(1989); Gifu, Japan: Mizoguchi et al. (1996); Nagano, Japan: Takada Russia: Bromlei (1973). (1979); b - FO frequencyof occurrencestandardized to 100%;RV - relative volume. 120 Ursus 13:2002 size (10-20 kg) and apparenthigh density of ungulates 1988, Schalleret al. 1989). Often, several scats occurred (muntjacs and serows) in our study area may have nearby,suggesting that they were sites of extended rest promptedbears to be morecarnivorous than reported else- periods. Ground nests of Asiatic black bears in China where. Bears may encounterthese ungulatesespecially appearedsimilar to nests thatwe observedin Taiwanmade frequentlyin oak areas, where bears and ungulatesboth by wild boar. In contrast,bears in Taiwantended to bend congregate. Withvarying availability of fruits,it may be and twist tall grassor shrubsand formthem into a distinc- that abundantpopulations of small-bodiedungulates are tive bowl, ratherthan bite off and pile up pieces of grass importantfor sustaining black bear populations in Tai- or broken branches. Similar bowl-shaped ground nests wan and otherparts of Asia. made of bamboowere apparentlyconstructed by bearsin Arborealfeeding seems to be common among Asiatic South Korea (W.M. Kim, National Instituteof Environ- black bears in broadleafforests (Schaller 1969, Bromlei mental Research, Incheon, South Korea, personal com- 1973, Wang 1988, Saberwal 1989, Schaller et al. 1989, munication, 2001). Some hunters that we surveyed Reid et al. 1991, Ma et al. 1994). The large numberof thoughtthat groundnests were made for securityor prey reportsof this behaviorsuggests that Asiatic black bears surveillance. Because these huntershad experienceswith are morearboreal than American black bears; this may be bearsconsuming their trap-captured prey, it is perhapsex- partly explained by differences in the structureof their pected that they would suggest a predatoryexplanation feet (Pocock 1932; Huygens et al. 2000, unpublishedre- for the nests. port)as well as a greaterprevalence of tree-borneinstead Fromthe informationat hand,it is difficultto ascribea of a bush-bore fruitsin Asian forests. AlthoughAmeri- reason for these groundnests. We posit 3 different,but can blackbears build tree nests in some areas(D. Garshelis, non-mutuallyexclusive explanations. (1) Because we personal observation), such nest building seems much found several such nests near our field camp after being more common amongAsiatic black bears(Bromlei 1973, raidedby bears, we presumedthat bears used these nests Schaller 1969, Nozaki et al. 1983, Schaller et al. 1989, while they remainedin the area. Possibly the function,in Reid et al. 1991, Ma et al. 1994, Huygens et al. 2000, this case, was to provide protectivecover when near hu- unpublishedreport). It is also common among sun bears man threats, similar to what appearsto be the case for (Helarctos malayanus; Erdbrink1953, McConkey and some tree nests built by sun bears and Andean bears Galetti 1999, Meijaard1999), which inhabitmany of the (Goldstein1991, 2002; Meijaard1999). (2) Groundnests same areasas Asiatic black bears. also tended to be built along rocky ridges, near cliffs, or It is not entirely clear whether tree nests serve a spe- beside trails,where terrainwas uneven, and so may have cific purposeor merely arise as a result of feeding activ- functionedmainly to enhanceresting comfort. (3) Ridges, ity. Whereasthese platformsmay be a comfortableand cliffs, and trailsalso tend to funnel the movementsof un- secure place for a bear to sit, our observations suggest gulates, which in this populationserved as a food source that in Taiwanthey are mainly the result of bears pulling for bears. Hence, it may be that the nests were, as the food-ladenbranches toward the relativelystable center of hunterspresumed, used by bears to ambushprey, espe- the tree. Most of the platforms we saw consisted of a cially where other groundcover was lacking. It may not nest- jumbledarray of branches,rather than a carefullyformed havebeen coincidentalthat the mostelaborate ground structure. Indigenoushunters indicated that they rarely building so far observed for this (or any other bear) spe- saw bears resting in these platforms. Observationsby cies was in an areawhere ungulate prey was an important others, however, suggest that Asiatic black bears may dietarycomponent. sometimes construct platforms for rest and protection, especially when the ground is water-logged (Yin 1954, Comparisonsof Methodologies Steinmetzet al. 1999). Andeanbears (Tremarctos ornatus) We used 3 methods to investigate bear diets. Each of relianceon certain may use tree nests for resting or to guard feeding sites, the methodsindicated a strong species soft mast in summer especially human-relatedfoods (Goldstein 1991, 2002); of hard mast in autumnand (Table items thatwere not similarly,sun bear tree nests seem most common in hu- 1), but each also yielded some unique man-disturbedareas (Meijaard1999). identifiedby the othermethods (Table4). Groundnests of black bears have been remarkedupon Putnam (1984:79) observed that scats are "the most source of informa- much less frequentlythan tree nests. Asiatic black bears readily-availableand easily-collected have been observed to rake leaves into a bed for resting tion"for assessing diets. The principaldifficulty in inter- reflect amounts duringcold or rainy weather (Bromlei 1973, Wu 1983). pretingscats is thatthey only proportional not foods consumed and Largebeds (130 cm diameter,3-30 cm deep) constructed of undigestedmaterial, (Hewitt We faced an additional of piled up broken bamboo, shrubs, or vines have been Robbins 1996, Litvaitis 2000). Putman's in reportedfor Asiatic blackbears in China(Wu 1983, Wang problem,notwithstanding (1984) statement, DIETS OF ASIATICBLACK BEARS * Hwang et al. 121

Table4. Attributesof 3 methodsused to obtaindietary information on Asiaticblack bears in YushanNational Park, Taiwan, 1998-2000.Our subjective scoring was basedon the followingscale: -virtually no information;+little; ++ moderate; +++ high. Scats Feeding sign Interviews Foods eaten (identified items; unique items)a + (19; 5) ++ (26; 8) +++ (70; 48) Plant partsconsumed +++ + ++ Volume consumed ++ + + Foragingbehavior -+ ++ Seasonalityof diet + + ++ Reliability of data ++ + + Efficiency of data collection + ++ +++ a Total numberof food items identifiedusing particularmethod; number of items uniquely identifiedusing that method (n = 84 items identified among all 3 methods). the mere collection of scats. Our sample in the spring we made no attemptto quantitativelycompare feeding was so small thatwe could not reliablyascertain the spring sign by season, or even with other methods, but used the diet from scats. A relatively small sample size of scats in data instead to increase our knowledge of what and how spring or summer also was reported in other studies bears ate. (Bromlei 1973, Takada1979, Manjrekar1989, Schaller We identifiedmore individual food items from sign than et al. 1989, Hashimoto and Takatsuki1997). A paucity from scats (Table4) with roughlythe same overall sample of scats in springmay be due to low ingestionrates (Roth size (i.e., total n > 600, Table 1); hence, sign seemed a 1980, Mattsonet al. 1991) and high rates of decomposi- more efficient means of data collection (Table4). Addi- tion of scats composed of green vegetation, particularly tionally, more behavioralinformation was gained from in humid environmentssuch as Taiwan. Additionally,in sign (e.g., tree and ground nests) than scats. Sign also spring and summer,black bears in Yushanwere highly persisted longer, so gave a better record of past feeding dispersed (Wang and Hwang 1999, 2000), making scats events (i.e., 1-2 yearsbefore), althoughthat made it more difficult to find. Conversely,in autumn,scats composed difficultto determinethe seasonof the feedingevent. Fur- of acornswere found in abundancebecause bears congre- thermore,in some cases we could not identify from feed- gated in a feeding areaand likely consumedlarge quanti- ing sign the actualfood that was eaten. ties of abundant food and thus defecated frequently. Interviewswith indigenous huntersyielded a long list Moreover, acorn scats persisted for 1-2 months in dry of bearfoods and informationabout when and how bears oak forests. The disproportionatenumbers of scats found ate them (Table 4). Other studies have similarly docu- duringthe 3 seasons is thus a substantialsource of bias. mented the striking breadthand depth of "what people Anothersource of bias relatesto wherescats were found. untutoredin science, know about plants and animals" In spring and summerwe collected scats while hiking on (Ellen 1998:89). Ellen (1993) found that the Nuaula trails and radiotracking;thus we were more apt to find people in southeastAsia knew an enormousamount about scats within the elevational range that was most acces- foods and behaviors of wild pigs, an animal that they sible to us. We spent most of the fall trappingin an oak- hunted. Hunters'knowledge aboutbears in Taiwancame rich area that attractedbears, so we found many scats from personal observations and experiences conveyed composed largely of oak mast. We sometimes found a throughgenerations; thus, it covered a much longer pe- large group of scats in the same general area,or even un- riod and broaderarea than we observedfirst hand,and so der the same tree. Althoughit is likely that these did not was less prone to spatialor temporalbiases. These data representindependent sample units, we had no way of were clearly more inclusive of springand summerfoods distinguishing true sample units (scats from different than data obtained from scats and feeding sign. How- bears). ever, food items listed by interviewees were difficult to Feeding sign had some of the same biases as scats, quantifyin terms of relative importanceto bears. As an namely non-independenceof individual trees that occur approximate index to this, we used the number of in clumps and a possibly non-representativesampling re- intervieweesreporting each food item (Table 1). flective of our activities. Additionally,bears leave evi- The principalproblem with collecting ethnobiological dent sign when feeding on some types of foods but not informationin our study was the interpretationand reli- others. Herbaceousplants are absentin our dataon feed- ability of the data (Table4). Interpretationsmay be con- ing sign, as would be some shrub-bornefruits and certain foundedby languagedifficulties, as some ethnobiologists types of insects. Some trees show clear evidence of bear have indicated (Posey 1992, Ellen 1998, Berkes 1999). feeding, especially species in which bearsbreak branches, We attemptedto avoid this problem by living in the vil- whereas others may be more difficult to detect. Hence, lages, going into the field with interviewees, and collect- 122 Ursus 13:2002 ing samplesof food items thatwere discussed. However, sistants,equipment, and logistical support. Camperland informantssometimes used local names thatwe could not InternationalLtd. donatedoutdoor gear. Dr. G. Wangof translateor describedplants that they could not name. The NationalTaiwan Normal University helped identifyplant largerpotential problem was misinformationrelayed to specimens, and Mr. W. Lin and D. Changof YushanNa- us becauseof misconceptionsby the local hunters. These tionalPark assisted with translationsof local plantnames. huntershad a vested interestin being astute observersof We are gratefulto the many volunteerswho helped with naturebecause they had to rely upon their knowledge to fieldwork and the indigenous villagers who generously obtain food for themselves and theirfamily (Ellen 1998, contributedvaluable informationduring our interviews. 1999; Berkes 1999; Berlin 1999; Diamond and Bishop D.G. Reid and F.T.van Manen providedinsightful com- 1999). However, most huntersdid not specifically seek ments on an earlierdraft of this manuscript. bears, and the circumstancesin which they most often encounteredbears or their sign, namely aroundtheir un- gulate trapsor sometimes in their crop fields, may have LITERATURECITED AND H. TORRI. of prejudicedtheir impressions. Huntersseemed well aware AZUMA,S., 1980. Impact humanactivities on thatbear diets were invari- survivalof the Japaneseblack bear. InternationalConference basically vegetarian,yet they on Bear Researchand 4:71-79. listed as a common food item even Management ably ungulates though BALLARD,W.B. 1992. Bear on moose: a review of few huntersever saw bears We do predation attackingungulates. recent North American studies and their management know how basedtheir of bear not many opinions carivory implications. Alces Supplement1:162-176. on scats or naturalsign (carcasses)versus just bearsign at BEEMAN,L.E., ANDM.R. PELTON.1980. Seasonal foods and theirtraps. We also wonderedat times whethertheir de- feeding ecology of black bears in the Smoky Mountains. scriptionsof bear diets were based only on direct obser- InternationalConference on Bear Research and Management vations or included some degree of supposition, 4:141-147. F. 1999. Sacred traditional imagination, and enhancement (Peyton 1980, Berkes BERKES, ecology: ecological knowledgeand resource management. Taylor and Francis 1999, Huntington2000). We were aware of one case Press, Philadelphia,Pennsylvania, USA. where informationwas evidently obtained from televi- BERLIN,B. 1999. How a folkbotanicalsystem can be bothnatural sion because it concernedanother of bear. species andcomprehensive: one MayaIndian's view of the plant None of the 3 methods that we used to ascertain the world. Pages71-89 in D.L. Medinand S. Atran,editors. diet of black bears was unbiasedor error-free. They had Folkbiology.MIT Press, Cambridge, Massachusetts, USA. differingattributes (Table 4), however,which made them BIERNACKI,P., ANDD. WALDORF.1981. Snowballsampling: complimentaryrather than redundant. We recognize that problem and techniques of chain referral sampling. in most otherstudies, scats are sufficientto describebear Sociological Methods and Research 10:141-163. ANDM.R. PELTON. 1988. Seasonal in diets. More attention to feeding sign, however, might BRODY,A.J., changes of behaviors. in digestion in black bears. Canadian Journal of Zoology improveinterpretations feeding Finally, 66:1482-1484. situations such as where bears are at low ours, density BROMLEI,F.G. 1973. Bearsof the southfar-eastern USSR. andwet conditionscause scats to rapidlydisintegrate, tra- IndianNational Scientific Documentation Centre, New Delhi, ditional ecological knowledge may greatly enhance the India (Translated from Russian; originally published by biological data set. To biologists wary of second-hand Nauka, Moscow-Leningrad,USSR, 1965). information,we suggestthat time spentwith local, knowl- BUNNELL,F.L., ANDT. HAMILTON.1983. Forage digestibility edgeablepeople may be more productivethan time spent and fitnessin grizzlybears. InternationalConference on BearResearch and 5:179-185. wanderingabout the forest looking for scats and sign of Management rareanimals. CHEN,J-S. 1991. A study on the ecology of Asiatic black bear. Abstractsof the East Asiatic Bear Conference2:147. CHEN,S-J. 1989. Geological surveyof San-chaMountain areas ACKNOWLEDGMENTS in YushanNational Park. YushanNational Park, Nantou, Taiwan. This was fundedmainly by YushanNational Park, (In Chinese.) study Y-F. 1989. The of the Nan-shi Taiwan. The Park also researchassistants and CHEN, vegetationsurvey forestry provided roadin YushanNational Park. Yushan National Park, Nantou, with field Additional helped logistics. monetarysupport Taiwan. (In Chinese.) of Minnesota the was provided at the University by DEBRUYN,T.D. 1997. Habitatuse, food habit and population MacArthurProgram on Peace, Justiceand Sustainability; characteristicsof female black bears in the central Upper the ConservationBiology Program;the James W. Wilkie Peninsula of Michigan: A geographicinformation system Natural History Fund; and the GraduateSchool. The approach.Dissertation, Michigan Technological University, Wildlife Laboratoryof the Departmentof Biology, Na- Houghton,Michigan, USA. of the tional Taiwan Normal University provided researchas- DIAMOND,J., ANDK.D. BISHOP.1999. Ethno-omithology DIETSOF ASIATIC BLACK BEARS * Hwang et al. 123

Ketengbanpeople, IndonesianNew Guinea. Pages 17-45 americanus,food habitsin southwesternAlberta. Canadian in D.L. Medin and S. Atran, editors. Folkbiology. MIT Field-Naturalist105:335-345. Press, Cambridge,Massachusetts, USA. HORINO,S., ANDS. MIURA.2000. Populationviability analysis EAGLET.C., AND PELTON, M.R. 1983. Seasonalnutrition of black of a Japaneseblack bear population. Population Ecology bears in the Great Smoky Mountains National Park. 42:37-44. InternationalConference on BearResearch and Management HORIUCHI,M., S. HAYASHI,Y. YOSHIDA,T. TSUBOTA,AND T. 5:94-101. HAZUMI.2000. Food habits of Japaneseblack bear (Ursus EILER, J.H., W.G. WATHEN, AND M.R. PELTON. 1989. thibetanusjaponicus) as assessed by fecal analysis in Neo Reproductionin black bears in the southern Appalachian Village, Gifu Prefecture.Chubu Forestry Research 48:149- Mountains. Journalof Wildlife Management53:353-360. 152. (In Japanese.) ELLEN, R.F. 1993. Nuaulu Ethnozoology: a systematic HUNTINGTON,H.P 2000. Using traditionalecological knowledge inventory. Centrefor Social Anthropologyand Computing, in science: methodsand applications. Ecological Application in co-operationwith the Centrefor South-EastAsian Studies, 10:1270-1274. University of Kent, Canterbury,Kent, United Kingdom. HUYGENS,O.C., ANDH. HAYASHI.2001. Use of stone pine seeds 1998. Indigenous knowledge of the rainforest: andoak acornsby Asiatic blackbears in centralJapan. Ursus perception,extraction and conservation. Pages 87-99 in B.K. 12:47-50. Maloney,editor. Humanactivities and the tropicalrainforest: HWANG,M-H., ANDY. WANG.1993. A preliminaryobservation past, present and possible future. Kluwer Academic on a captiveFormosan black bear. TaipeiZoo Bulletin 5:71- Publisher,Dordrecht, Netherlands. 87. . 1999. Models of subsistence and ethnobiological ,ANDD.L. GARSHELIS.2000. Preliminarystudy knowledge: between extractionand cultivationin Southeast on activity patternsof Formosan black bears (Selenarctos Asia. Pages 91-117 in D.L. Medin and S. Atran, editors. thibetanusformosanus) in YushanNational Park, Taiwan. Folkbiology. MIT Press, Cambridge,Massachusetts, USA. Journalof NationalPark 10:26-40. (In Chinesewith English ERDBRINK,D.P. 1953. A review of fossil and recentbears of the abstract.) old world with remarkson theirphylogeny based upon their KASBOHM,J.W., M.R. VAUGHAN,AND J.G. KRAUS.1995. Food dentition. DrukkerijJan de Lange, Deventer,Netherlands. habits and nutrition of black bears during a gypsy moth FURUBAYASHI, K., K. HIRAI, K. IKEDA, AND T. MIZUGUCHI. 1980. infestation. CanadianJournal of Zoology 73:1771-1775. Relationships between occurrence of bear damage and KHRAMTSOV,V.S. 1997. On the feeding behavior of Asiatic clearcutting in central Honshu, Japan. International black bear. Byulleten' Moskovskogo Obshchestva Conferenceon Bear Researchand Management4:81-84. Ispytatelei Prirody Otdel Biologicheskii 102:39-41. (In GARSHELIS,D.L., AND M. R. PELTON. 1981. Movements of black Russian with English abstract.) bearsin the GreatSmoky MountainsNational Park. Journal KIRKPATRICK,R.L., ANDP.D. PEKINS.2002. Nutritionvalue of of Wildlife Management45:912-925. acornsfor wildlife. Pages 173-181 in W.J.McShea and W.M. GOLDSTEIN,I. 1991. Are spectacled bear's tree nests feeding Healy, editors. Oak forest ecosystems: ecology and platformsor resting places? Mammalia55:433-434. managementfor wildlife. Johns Hopkins University Press, . 2002. Spectacledbear-cattle interactions and tree nest Baltimore,Maryland, USA. use in Bolivia and Venezuela. Ursus 13:369-372. Kou, C-M. 1999. Studies on the ecological resources and GRABER,D.M., ANDM. WHITE.1983. Black bearfood habits in management of Warabi region in Yushan National Park. Yosemite National Park. InternationalConference on Bear YushanNational Park, Taiwan. (In Chinese with English Researchand Management5:1-10. abstract.) GRENFELL,W.E., JR., AND A.J. BRODY. 1983. Seasonal food of KUNKEL,K.E., ANDL.D. MECH.1994. Wolf and bear predation black bears in TahoeNational Forest, California. California on white-tailed deer fawns in northeastern Minnesota. Fish and Game 69:132-150. CanadianJournal of Zoology 72:1557-1565. HASHIMOTO,Y., AND S. TAKATSUKI.1997. Food habits of Japanese LANDERS, J.L., R.J. HAMILTON,A.S. JOHNSON,AND R.L. blackbears: a review. HonyuruiKagaku (Mammal Science) MARCHINTON.1979. Food and habitat of black bears in 37:1-19. (In Japanesewith English abstract.) southeastern North Carolina. Journal of Wildlife HAzuMI,T. 1994. Statusof Japaneseblack bear. International Management43:143-153. Conference on Bear Research and Management9(1):145- LEE,H., Z. HWANG,H. HWD,AND H. CHOO.1991. The bear's 148. food and mode of life. Abstractsof the East Asiatic Bear ,ANDN. MARUYAMA.1986. Movementsand home ranges Conference2:134. of Japaneseblack bears in Nikko. InternationalConference LEE,K.H., S. CHONG,H. MUN,S. KIM,AND J.S. CHUNG.1991. on Bear Researchand Management6:99-101. Ecology of Bandhalbear. Abstractsof the East Asiatic Bear , AND . 1987. Movements and habitat use of Conference2:134-135. black bearsin Japanese Nikko. InternationalConference on LITVAITIS,J.A. 2000. Investigating food habits of terrestrial Bear Researchand Management7:275-279. vertebrates. Pages 21-34 in L.B. Boitani and T.K. Fuller, AND C.T. ROBBINS. 1996. HEWITT,D.G., Estimating grizzly bear editors. Researchtechniques in animal ecology. Columbia food habits from fecal analysis. Wildlife Society Bulletin UniversityPress, New York,New York,USA. 24:547-550. MA,Y.Q, J.C. Hu, ANDQ.L. Di. 1994. Bears in China. Sichuan HOLCROFT, ANDS. HERRERO. A.C., 1991. Black bear, Ursus Science PublicationHouse, Sichuan, China. (In Chinese.) 124 Ursus 13:2002

MANJREKAR,N. 1989. Feeding ecology of the Himalayanblack York,New York,USA. bear (Selenarctosthibetanus Cuvier) in DachigamNational PRITCHARD, G.T., AND C.T. ROBBINS. 1990. Digestive and Park. Thesis, SaurashtraUniversity, Rajkot, India. metabolicefficiencies of grizzly and black bears. Canadian MATHEWS,N.E., AND W.F. PORTER. 1988. Black bear predation Journalof Zoology 68:1645-1651. of white-tailed deer neonates in the central Adirondacks. PUTMAN,R.J. 1984. Factsfrom faeces. MammalReview 14:79- CanadianJournal of Zoology 66:1241-1242. 97. MATTHEWS,S.J. 1977. The seasonal,altitudinal, and vegetational RAINE, R.M, AND J.L. KANSAS. 1990. Black bear seasonal food incidence of black bear scats in the GreatSmoky Mountains habits and distributionby elevation in Banff NationalPark, NationalPark. Thesis, Universityof Tennessee, Knoxville, Alberta. InternationalConference on Bear Research and Tennessee, USA. Management8:297-304. MArrsON,D.J. 1998. Diet andmorphology of extantand recently REID, D., M.D. JIANG, Q.T. TENG, Z.S. QIN, AND J.C. Hu. 1991. extinct northernbears. Ursus 10:479-496. Ecology of the Asiatic black bear (Ursus thibetanus) in , B.M. BLANCHARD,AND R.R. KNIGHT.1991. Food habits Sichuan,China. Mammalia55:221-237. of Yellowstonegrizzly bears, 1977-1987. CanadianJournal RoTH,H.U. 1980. Defecation rate of captive brown bears. of Zoology 69:1619-1629. InternationalConference on BearResearch and Management MCCONKEY K., AND G. GALETTI. 1999. Seed dispersal by the 4:249-253. sun bearHelarctos malayanus in centralBorneo. Journalof SABERWAL,V. 1989. Distributionand movementpatterns of the TropicalEcology 15:237-241. Himalayan black bear (Selenarctos thibetanus Cuvier) in MCDONALD, J.E., JR., D.P. FULLER, T.K, FULLER, AND J.E. Dachigam National Park, Kashmir. Thesis, Saurashtra CARDOZA.1994. The influenceof food abundanceon success University,Rajkot, India. of Massachusettsblack bear hunters. Northeast Wildlife SCHALLER,G.B. 1969. Food habits of the Himalayan black 51:55-60. bear (Selenarctos thibetanus)in the Dachigam Sanctuary, , ANDT.K. FULLER. 1993. Blackbear food habits:beyond Kashmir. Journal of Bombay Natural History Society the same old scats. InternationalUnion of Game Biologists 66:156-159. Congress 21(1):293-288. , Q.T TENG,K.G. JOHNSON,X.M. WANG, H.M. SHEN, MELAARD,E. 1999. Ursus(Helarctos) malayanus, the neglected ANDJ.C. Hu. 1989. The feeding ecology of giant pandas Malayansun bear. NetherlandsCommission for International and Asiatic black bears in the Tangjiahe reserve, China. Nature Protection, Backhuys Publishers, Leiden, Pages 212-241 in J.L.Gittleman, editor. Carnivorebehavior, Netherlands. ecology, andevolution. CornellUniversity Press, New York, MIZOGUCHI,N., A. KATAYAMA,T. TSUBOTA, AND A. KOMIYAMA. New York,USA. 1996. Effect of yearly fluctuationin beechnut production SCHWARTZ,C.C., ANDA.W. FRANZMANN.1991. Interrelationships on food habits of Japaneseblack bears. HonyuruiKagaku of blackbears to moose andforest succession in the northern 36:33-44. (In Japanesewith English abstract.) coniferous forest. Wildlife Monographs113. NAGANAWA,K., ANDM. KOYAMA.1994. Food habits of Japanese STEINMETZ,R., T. STONE,AND T. CHAN-ART. 1999. An ecological black bear in Tanzawamountain area. Transactionsof the survey of habitat, wildlife and people in Xe Sap National Lao PDR. JapaneseForestry Society 105:539-542. (In Japanese.) BiodiversityConservation Area, Saravan province, NELSON, R.A., G.E. FOLK, JR., E.W. PFEIFFER,J.J. CRAIGHEAD, WWF for Nature-ThailandProgramme Office, Bangkok, C.J. JONKEL, AND D.L. STEIGER. 1983. Behavior, Thailand. black bears biochemistry,and hibernationin black, grizzly, and polar TAKADA,Y. 1979. The food habits of the Japanese bears. International Conference on Bear Research and in the centralmountain area, Nagano prefecture. Journalof of 8:40-53. Management5:284-290. the Mammalogical Society Japanese (In NOZAKI,E., S. AZUMA,T. Aoi, H. TORII,T. ITO,AND K. MAEDA. Japanesewith English abstract.) 1983. Food habits of Japanese black bear. International TALBOT,S.L., ANDG.F. SHIELDS. 1996. A phylogeny of the bears Conferenceon Bear Researchand Management5:106-109. (Ursidae) inferred from complete sequences of three and PARK,S. 2001. Habitat-basedpopulation viability analysis for mitochondrial genes. Molecular Phylogenetics the Asiatic black bear in Mt. Chiri Mational Park, Korea. Evolution 5:567-575. of black bears Centrum for Biologisk Mangfald (CBM:s) Skriftserie TERNENT,M.A. 1995. Managementand ecology Minnesota. Thesis, (Uppsala) 3:149-165. in Camp Ripley Military Reservation, USA. PEYTON,B. 1980. Ecology, distribution,and food habitatsof University of Minnesota,Minneapolis, Minnesota, Food habitatsof the black bear in the spectacled bears, Tremarctosornatus, in Peru. Journalof TORI,H. 1989. Japanese Ohwi River. Journalof Mammalogy61:639-652. headwatersof the JapaneseForestry with POCOCK,R.I. 1932. The black and brown bears of Europeand Society 71:417-420. (In Japanese English abstract.) and bears. Asia. Journalof Bombay NaturalHistory Society 36:101- VAUGHAN,M.R. 2002. Oak trees, acorns, Pages 138. 224-240 in W.J.McShea and W.M. Healy, editors. Oakforest for wildlife. Johns POSEY,D.A. 1992. Interpretingand applying the "reality"of ecosystems: ecology and management Baltimore, USA. indigenous concepts: what is necessary to learn from the Hopkins UniversityPress, Maryland, of black bears. Chinese Wildlife natives? Pages 21-34 in K.H. Redford and C. Padoch, WANG,C.M. 1988. Ecology editors. Conservationof Neotropicalforest: working from 42:16-18. (In Chinese.) traditionalresource use. Columbia University Press, New WANG,Y. 1999. Statusand management of the Formosanblack DIETS OF ASIATICBLACK BEARS * Hwang et al. 125 bear in Taiwan. Pages 213-215 in C. Servheen,C. Herrero, English abstract.) andB. Peyton,editors. Bears: statussurvey and conservation , H-S. CHEN, AND M-F. KAO. 1992. The action plan. IUCN, World Conservation Union, Gland, ecological study and managementof Formosanblack bear Switzerland. (III): case studyon an introducedbear. Ecological Research , H-S. CHENG, AND J-L. HWANG.1993. The ecological Report,Council of Agriculture,Taipei, Taiwan. (In Chinese.) study and management of the Formosan black bear WELCH,C.A., J. KEAY, K.C. KENDALL,AND C.T. ROBBINS.1997. (Selenarctos thibetanus formosanus) (IV). Council of Constraintson frugivoryby bears. Ecology 78:1105-1119. Agriculture,Taipei, Taiwan. (In Chinese.) WILTON,M.L. 1983. Black bearpredation on young cervids- ,S-W. CHU,AND U.S. SEAL, EDITORS.1994. Asiatic black a summary. Alces 19:136-146. bearpopulation and habitat viability assessment. TaipeiZoo, Wu, J-Y. 1983. Preliminary observations on the food Taipei,Taiwan, and Conservation Breeding Specialist Group, specializationof black bearsfrom Qinling. Chinese Journal Apple Valley, Minnesota,USA. of Zoology 4:48-51. (In Chinese.) , AND M-H. HWANG. 1999. Ecology of Formosan black YIN,U.T. 1954. Habits of the Asiatic black bear (Selenarctos bears and bear-people interactionin YushanNational Park thibetanusthibetanus G. Cuvier). Journalof BombayNatural (I). Yushan National Park, Nantou, Taiwan. (In Chinese History Society 52:586-587. with English abstract.) Received: 28 , AND . 2000. Ecology of Formosan black bears May 2001. and bear-people interactionin YushanNational Park (II). Accepted: 14 March 2002. YushanNational Park, Nantou, Taiwan. (In Chinese with Associate Editor: van Manen.