sign in sign up
<<
Home , Leopardus, Spectacled bear

Andean density and abundance estimates — How reliable and useful are they?

David L. Garshelis1

Minnesota Department of Natural Resources, Grand Rapids, MN 55744, USA

Abstract: Few attempts have been made to estimate numbers and densities of Andean ( ornatus). It is understandable that the many challenges involved in these efforts have made it difficult to produce rigorous estimates. A crude estimate of ,20,000 Andean bears was derived by extrapolating the lowest observed density of American black bears ( americanus) across the range of Andean bears. A second estimate, based on rangewide genetic diversity, produced a wide range of values; however, the low end of the confidence interval roughly matched the population estimate based on minimum black bear density. A mark– recapture analysis of 3 camera-trapped bears in also yielded a similar density (4.4–6 bears/100 km2), but overlapping home ranges of 2 radiocollared bears at that same site suggested a higher density (.12 bears/100 km2). Neither of these estimates can be considered reliable or representative of the wider population because of the small sample sizes. Moreover, the effective sampling area for the camera-trapping study was uncertain. A DNA hair-trapping mark–recapture study in sampled a greater number of bears (n 5 25) within a larger study area, but a male-biased sex ratio suggested that closure was violated, precluding a simple estimate of density based on the area of the trapping grid. Also, low capture rates in what was perceived (from incidence of bear sign) as prime bear habitat might be indicative of a sampling bias. These issues are not simply incorporated into confidence intervals (CIs): CIs only include uncertainty due to sampling error, not biased sampling or an ambiguous sampling area. Whereas these (low density) estimates may provide guidance for conservation, their greatest usefulness may be in providing directions for improvement of future studies of Andean bears, as well as bears in Asia, which also lack rigorous population estimates.

Key words: biased estimates, camera-trapping, DNA hair-sampling, effective trapping area, genetic diversity, mark–recapture, precision, radio-telemetry, representative sampling, Tremarctos ornatus

Ursus 22(1):47–64 (2011)

The pursuit of estimates of abundance is Vaughan 2009, Clark et al. 2010), to extrapolate pervasive, partly because such estimates are useful densities from small study sites to larger areas for management and conservation, and partly (Miller et al. 1997, Boulanger et al. 2002, Robinson because they satisfy a certain basic human intrigue. et al. 2009) or to apply across large regions In North American bear studies, population esti- (Bellemain et al. 2005, Garshelis and Noyce 2006, mates are generally derived from mark–recapture Dreher et al. 2007, Kendall et al. 2009). sampling approaches, either involving actual bear For a variety of reasons, these rigorous population captures or ‘‘recaptures’’ obtained from photo- estimation techniques have rarely been applied to graphs, sightings, or DNA from hair samples. Much Andean bears (Tremarctos ornatus) in South Amer- effort has been and is being invested in developing ica (or, for that matter, to the bears in Asia, other scientifically rigorous population estimation tech- than giant pandas [ melanoleuca]). Logis- niques, with associated confidence intervals, that can tical and financial constraints are often prohibitive, be used to guide management of bears on small, bear densities may be too low to obtain adequate selected areas (Gervasi et al. 2008, Immell and samples, and the motivation to obtain such estimates Anthony 2008, Obbard and Howe 2008, Tredick and may be lacking because the bears are not legally hunted. Nevertheless, there has been growing inter- [email protected] est as well as a few recent attempts to obtain

47 48 ANDEAN BEAR POPULATION ESTIMATES N Garshelis abundance estimates of Andean bears using some densities ranged from 7–130/100 km2, with a median newly developed tools. of 25 bears/100 km2. Applying this median density to Here I review the current knowledge of Andean Andean bear range produced an estimate of 65,000 bear abundance. Rather than simply summarizing bears. Applying the lowest the few reports of population estimates, I describe density yielded an estimate of ,18,000 (0.07 x and thoroughly critique the methods and data. These 260,000). Including cubs increases the estimate to criticisms are not meant to rebuke the efforts of any .20,000. of the investigators, all of whom worked hard to The extreme lowest density reported for American collect and analyze data, but to help evaluate the black bears is from an area of marginal black bear veracity of the results and thereby aid in moving the habitat in Alaska — one of the northernmost black science of population estimation forward, especially bear populations ever studied (Miller et al. 1987, in the context of the Andean bear. Investigators are Miller 1994). This population was also hunted. Thus, often well aware of the foibles of their work, but may Peyton et al. (1998) were confident that the ,20,000 be dissuaded from highlighting those in the publica- estimate for Andean bears, derived from this very tion process. Here, I take the role of a post-hoc low black bear estimate, should be a minimal value. reviewer, aiming to enhance interpretations of past Moreover, the median American black bear density works and provide guidance for future efforts. was .3x higher even though a large number of the black bear density estimates were from hunted populations, whereas Andean bear populations are Rangewide population estimate derived technically unhunted (although certainly some ani- from density of a surrogate mals are illegally killed, especially when they While preparing a conservation action plan for depredate or crops; Jorgenson and Andean bears across their range, Peyton (1999:160) Sandoval-A 2005, Goldstein et al. 2006). summarized their status and distribution. He indi- cated that a group of experts were ‘‘confident that there are at least 18,250 wild spectacled [Andean] Rangewide population estimate derived bears.’’ These bears occupied at least 50 distinct from genetic diversity habitat fragments totaling 260,000 km2. Four of The degree of genetic variation in a population is a these fragments were estimated to contain .1,000 reflection of mutation rate, genetic drift, natural adult bears each. The source for these estimates of selection, immigration, and population size. A bear numbers was mentioned in an earlier report by positive relationship between genetic diversity and Peyton et al. (1998:87). The estimated total area population size has been corroborated with studies occupied was derived by ‘‘drawing boundaries on from a host of species (Frankham 1996), although topographic maps around areas where field expedi- the exact form of this relationship is more concep- tions confirmed the presence of bears.’’ These tual than empirical. In theory, however, with data on boundaries were based on habitats where Andean genetic diversity and assumed values for the other bears were known to exist. Bear abundance within parameters, one can calculate a long-term average this area was then estimated by extrapolating effective population size (Ne). Ne is the size of an densities of American black bears (Ursus americanus) idealized population that would produce the genetic to the geographic range of Andean bears. structure actually observed. Whereas this conceptual American black bears were used as a surrogate for definition is straightforward, the estimation of Ne Andean bears because (1) no empirical estimates of and its conversion to number of (N) has abundance existed for Andean bears at that time, (2) been performed in a number of ways, yielding varied at least 24 mark–recapture-based estimates were results (Luikart et al. 2010). Models are also available for American black bears (summarized by available to detect previous population bottlenecks Garshelis 1994), and (3) American black bears and from allelic frequencies. These procedures have Andean bears are primarily forest-dwelling omnivo- recently been employed, for example, to estimate rous ursids of similar size (although we now know the size of ancestral (Ursus arctos) that the 2 species are fairly different ecologically). populations (Saarma et al. 2007) and to quantify and Excluding cubs of the , American black bear date the collapse of giant pandas, other ,

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 49 and great apes (Aspi et al. 2006, Goossens et al. tended to center around ,0.1 N; few valid (data rich) 2006, Zhang et al. 2007). estimates of Ne were .0.5 N. Harris and Allendorf Ruiz-Garcia (2003) used these genetic approaches (1989) estimated a Ne /N ratio of 0.24–0.32 for grizzly to produce a crude estimate of population size and (brown) bears; this range encompasses the narrower change for Andean bears. He obtained genetic range calculated by Miller and Waits (2003) for the samples (hairs, bones, teeth, skin, and blood) Yellowstone population (0.27–0.32). collected by field researchers and from zoos in the Applying Harris and Allendorf’s range of Ne /N northern part of the bear’s range: , ratios to Ruiz-Garcia’s (2003) estimates of Ne would , and Ecuador (although most zoo samples yield total population estimates of 17,000–125,000 were of unknown origin). He analyzed the genetic Andean bears. Ruiz-Garcia (2003) asserted, howev- diversity of these samples, and using population- er, that a more credible range of Ne values was only genetic models concluded that (1) the species was 19,000–24,000, and also elected to use a much higher genetically impoverished, (2) populations in Vene- correction factor (0.75) to convert Ne to N. zuela, northern Colombia, and Ecuador were - Hence, starting with samples from parts of 3 ically isolated, but (3) these conditions dated back to countries and initially applying 2 genetic models, a ancestral Andean bears. Using estimated mutation range of plausible mutation rates, and a range of rates, he further calculated that these populations proportions to convert Ne to N, Ruiz-Garcia diverged some 24,000 ago, and thus the genetic (2003:91) ultimately honed in on a value of about fragmentation was not a result of the European 25,000 Andean bears rangewide. He rejected higher human invasion 500 years ago. calculated values because they were inconsistent with A major finding of this study was that populations ‘‘the IUCN census,’’ by which he meant the estimate of Andean bears did not pass through a genetic of Peyton (1999), derived from application of the bottleneck: Ruiz-Garcia (2003) concluded that pop- lowest density American black bear population to ulation size has remained constant over the evolu- the rangewide area of Andean bears. tionary history of this species. Applying a range of It must be stressed that the estimates stemming mutation rates varying by more than an order of from genetic diversity are averaged across the span magnitude, he calculated long-term effective popu- of time that the species has existed, not a recent point lation sizes. Two different mutational models pro- in time. Thus, if the species declined appreciably in duced widely varying results. For example, in one modern times, the long-term average would be model he obtained Ne estimates of 644–3,044 for essentially meaningless for a current assessment. Colombia and 325–1,348 for Ecuador. In the second Ruiz-Garcia (2003) concluded, though, that no model, his estimates were 1,888–12,875 for Colombia significant bottleneck occurred, so the long-term and 1,429–8,017 for Ecuador. estimate is essentially the same as the current Country-specific estimates of Ne were then ex- estimate. Given that, he argued that estimates in panded to rangewide estimates. Unfortunately, the the range of 100,000 bears, resulting from some process for doing so was not explained, so cannot be models, mutation rates, and Ne /N correction factors, evaluated here. The full span of reported estimates must be wrong owing to their discordance with the was 5,500–30,000 bears; it is unclear, though, how ‘‘census.’’ What seems clear is that Peyton’s (1999) the upper end of this range was calculated, given that estimate constrained the range of results reported the maximum values for just Colombia and Ecuador from the genetic-based estimate; thus, the 2 estimates totaled nearly 21,000, while and Bolivia are are not truly independent. Both started with a wide reported to have the most bears (69% of area of bear range, but were trimmed to a value near 20,000 range; Peyton 1999). bears. To be of conservation value, the estimates of More recent genetic studies, with larger sample effective population size (which is a function of sizes from across a larger geographic area and population fluctuations, sex ratio, proportion of including more microsatellite loci (Ruiz-Garcia et animals breeding, litter size, generation time, etc.) al. 2005, Viteri and Waits 2009) have found higher must be converted to real population size. Many genetic diversity than originally reported, which estimates exist in the literature for the ratio of Ne to would convert to a higher estimate of the current N. Frankham (1995) compiled results from a bear population. Even these recent estimates of multitude of studies and found that estimates of Ne genetic diversity may be biased low because the

Ursus 22(1):47–64 (2011) 50 ANDEAN BEAR POPULATION ESTIMATES N Garshelis molecular markers employed were derived from estimated width of an average home range (or the other bear species, which are not closely related to radius of a circular home range) to account for the Andean bears (Ruiz-Garcia et al. 2005, Viteri and bears’ use of areas beyond the camera sites. This Waits 2009). Using markers that were not ascer- width was initially estimated based on the average tained from the target species may inhibit the ability maximum distance between each individual’s photo- to discern differences between some alleles (referred graphed locations. This yielded a buffer width of to as ascertainment bias). Consequently, it seems only 0.25 km around the camera-trap sites, and a likely that the original population estimates of Ruiz- density estimate of 13.6 bears/100 km2 (95% CI 5 8– Garcia (2003) will be substantially modified by 19/100 km2). This buffer strip was obviously too future derivations based on genetic diversity. small, as a home range radius of 0.25 km would equate to a home range area of only 0.2 km2. A previous study in this same area by Paisley Local population estimate based on (2001), with radiocollared bears, yielded home range camera captures estimates of .7km2, or an estimated radius of The Andean bear has highly individualistic color 1.5 km. One of these previously collared bears markings, which may include variable white or - appeared to have been recaptured at Rı´os-Uzeda et colored rings or partial rings around the eyes and al.’s (2007) camera traps. Using this home range streaks or speckled patches on the muzzle, chin, radius to buffer the camera-trap locations and define neck, and chest. Often these markings are bilaterally the effective trapping area (Fig. 1), Rı´os-Uzeda et al. 2 asymmetrical. Hence, photos taken by remote (2007) calculated a density of only 6 bears/100km 2 cameras of the face, neck, and chest of these bears (3.6–8.5/100 km ). Castellanos (2004, 2011) observed can be used to identify individuals. Others have larger home ranges in Ecuador: 15 km2 for females employed individual identifications based on cam- and 59 km2 for males. Applying Castellanos’ female era-trap photos to estimate populations of various home range radius to the camera-trap area produced species of striped or spotted carnivores or ungulates, a density estimate of 4.4 bears/100km2 (2.6–6.2/ via a ‘mark’–resight approach (Karanth 1995, 100km2). Although the sex of the photographed Karanth and Nichols 1998, Elkan 2003, Trolle and bears was unknown, Rı´os-Uzeda et al. (2007) elected Ke´ry 2003, Silver et al. 2004, Jackson et al. 2006, not to use Castellanos’ male home range estimates to Marnewick et al. 2008, Trolle et al. 2008, Gerber et buffer the trapping area and concluded that a density al. 2010). The ‘marking’ phase is just an initial photo of 4.4–6 bears/100km2 was most likely. of a naturally-marked animal; subsequent photos Much recent literature on a large number of represent resighting events. camera-trapped species has been concerned with the This method of population estimation was ex- estimation of effective trapping area. One study of a ploited by Rı´os-Uzeda et al. (2007) for Andean bears known population of ( pardus) in an area straddling 2 adjacent protected areas in concluded that a strip width of K the mean distance northern Bolivia. They selected 17 camera-trap between camera captures produced a reasonably locations based on presence of bear sign; habitat accurate density estimate (Balme et al. 2009). Other that bears rarely used was discounted in the search studies found that better results were obtained with for potential trap sites. Chosen sites included 9 in buffer strips equal to the average maximum distance pa´ramo and 8 in the forest. They set 2 cameras at between captures rather than K that distance each site (to obtain opposing views of bears). They (Parmenter et al. 2003, Sharma et al. 2010). Various did not use bait or scent to attract bears to the site. model-based approaches have also been proposed Cameras were checked weekly for a period of that use the spatial distribution of all capture data to 1 month, after which 7 photographs of 3 identifiable estimate the effective area of use (Efford et al. 2004a, bears were obtained. 2009; Gardner et al. 2009, 2010; Royle et al. 2009); Mark–recapture analysis indicated that only these these approaches are better supported mathemati- 3 bears occupied this area. To convert this estimate cally than the rather informal buffer strip calcula- of abundance to density, the researchers estimated tions. Still other authors have argued that home the effective sampling area. They employed a ranges need to be estimated independent of the commonly-used technique whereby the trapping area camera trap data because, especially when there are is increased by a buffer strip equal to K the few recaptures, the distances between them are

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 51

(Program DENSITY, Efford et al. 2004b) of the recapture data (DNA hair samples in this case). Tredick and Vaughan (2009) considered the latter estimate to be more accurate in this case because the telemetry data were too limited. Obbard et al. (2010) concluded that maximum-likelihood-based estima- tors of the effective trapping area derived from spatial distribution of recapture data (Efford et al. 2009) performed better than ad hoc boundary strips because they include heterogeneity in detection and also because they incorporate the variance of the area estimate as part of the variance of the density estimate. However, these sorts of modeling proce- dures require more data than are likely to be obtained in an Andean bear camera-trapping study. In the case of Rı´os-Uzeda et al.’s (2007) study, it is clear that the camera recapture data alone would have underestimated the effective trapping area. But even the telemetry data that were used to correct the Fig. 1. Camera-trapping sites (shown as camera effective area were subject to considerable error. In icons) used by Rı´ os-Uzeda et al. (2007) to obtain a Paisley’s (2001) study, the collared bears often mark–recapture estimate of the population size of wandered out of range of the telemetry equipment. Andean bears in northern Bolivia. Three individual Including areas where radio signals emanated but bears were photographically identified. These au- reliable triangulated locations could not be achieved thors used 3 varying-sized buffers, as shown (adapted from their figure), to define the effective caused the estimated home range size for 1 of 2 bears trapping area and thereby estimate bear density. The to nearly double. Yet even this was an underestimate inside-most buffer (darkest shading) was based on because radio signals were not audible on 16–20% of estimated home range radius derived from camera tracking attempts within the area that the bears were captures of the same individual at different sites. known to occupy. Two outer buffers were based on home ranges estimated from 2 radiotelemetry studies (Paisley If, as suggested by Paisley’s (2001) locational and 2001; Castellanos 2004, 2011). Innermost (white) presence–absence data, the real home ranges of these area was assumed to be trapped, but actually shows bears was 20 km2 or more, the total area encom- a sizeable void where bears would not have encoun- passed by Rı´os-Uzeda et al.’s (2007) camera sites was tered any traps. equivalent to about 1 home range. Based on empirical camera-trapping data with , Maffei invariably underestimates of actual movements and Noss (2008) recommended a minimum study (McCarthy et al. 2008). Soisalo and Cavalcanti area size of 4 home ranges. The smaller the study (2006) found that GPS-collared (Panthera area size relative to home range size, the less onca) moved nearly twice as far as indicated by their geographic closure (Garshelis 1992), and hence the recapture distances, even though camera traps were greater the influence of the chosen buffer area on the set in places with many GPS locations. The density estimate. In other words, the smaller the photographic data alone overestimated density of camera-trapping area, the greater the dependence on jaguars by .70%. For ocelots ( pardalis), radiotelemetry data for estimation of home range camera-trap based buffers yielded density estimates size and effective sampling area. that were more than twice that obtained using Small sampling areas are also apt to be non- telemetry-based estimates of the effective area representative of the larger regional area by chance (Dillon and Kelly 2008). Conversely, in a study of alone. For example, Matthews et al. (2008) sampled American black bears, Tredick and Vaughan (2009) 2 study sites in California with camera traps and found that the estimated density based on a observed American black bear densities 7x higher on telemetry-derived buffer strip was, on one study site, 1 site than the other. Indeed, 1 of the sites had one of about 2x the estimate derived from spatial analysis the highest densities ever reported for this species

Ursus 22(1):47–64 (2011) 52 ANDEAN BEAR POPULATION ESTIMATES N Garshelis

(133 bears/100 km2). The 2 areas had similar habitat is prone to be flawed, pointing to the need for characteristics and were only 4 km apart, but during independent telemetry data. Notably, Rı´os-Uzeda et the period of camera trapping (,6 weeks) the site al.’s (2007) estimate of $15,000 in expenses and 24 with high density had an abundance of blackberries person-weeks of field effort to conduct their study (Rubus discolor), which may have attracted bears. In did not include the expenses and effort also needed fact, there were only 3 bears on 1 site and 5 on the to obtain the telemetry data upon which they relied. other, so the dramatic disparity in apparent density A final issue related to camera-trapping is the was really attributable to a difference of just 2 bears. ability to discriminate individuals from photographs With such small numbers of animals, results are not that are often taken at different angles and distances likely to be meaningful and may be misleading in from the bears and under varying ambient condi- representing the larger surrounding area. tions. For Andean bears, the key is to clearly see the The number, spacing, and layout of the traps are markings of their face, and possibly their neck and other important considerations. Karanth and Ni- chest. Rı´os-Uzeda et al. (2007) used 2 cameras at chols (1998:2855) stressed that it is important ‘‘to each site, positioned opposite each other, to avert the ensure coverage of the entire area, without leaving potential problems of obtaining a photograph of holes or gaps that were sufficiently large to contain a only 1 side of the face and not being able to match ’s movements during the sampling period.’’ In that with a different photograph of the other side. Rı´os-Uzeda et al.’s (2007) study, a large gap They recommended that future studies use 3 cameras occurred in the middle of the trapping area to better discriminate individuals. Rı´os-Uzeda et al. (Fig. 1). This area was not trapped because it was (2007) claimed to be certain of individual identifica- considered marginal habitat for bears, and thus tion for each of the 7 bear photographs that they largely unused (R. Wallace, Wildlife Conservation obtained. However, a problem with photographs is Society, La Paz, Bolivia, personal communication, that identification of individuals is subjective; 2009). That seems like a logical decision in terms of investigators may differ in their judgments about trapping efficiency, but it is difficult to reconcile, distinguishing characteristics (Zug 2009). Had Rı´os- after the fact, whether this area should have been Uzeda et al. (2007) misidentified just 1 bear, their included in the density estimate. estimates of density would have been very different. Rı´os-Uzeda et al. (2007) demonstrated the poten- More photos of more bears would reduce the effect tial feasibility of employing camera-based individual of a small number of misidentifications, but would identifications for population estimation of Andean increase the chance of encountering many similar- bears. But the study also highlighted shortcomings looking bears; indeed, there may be a limit to the that might be improved in future efforts of this kind, number of individual bears that can be reliably depending on financial and logistical constraints distinguished from photographs, especially those (e.g., larger study area with more camera traps and taken in uncontrolled field conditions. no large gaps between traps). A major issue, though, that may be difficult to control is sample size; this depends both on bear density and their likelihood of Local population estimate based passing in front of a camera (which would be on radiotelemetry enhanced with a lure, but that also influences Paisley (2001) radiocollared and tracked 2 bears in recapture rates and may increase heterogeneity in essentially the same area where Rı´os-Uzeda et al. probability of capture). Camera trap studies of other (2007) later conducted their camera-trapping study. species have generated population estimates from Although Paisley (2001) did not attempt to estimate small sample sizes (e.g., ,6 individuals; McCarthy et density because of the small sample size, a crude al. 2008, Lynam et al. 2009, Rayan and Mohamad density estimate can be extracted from her data. The 2 2009), but in such cases individual capture hetero- bears were radiotracked mainly from ridgelines and geneity is likely to go undetected, resulting in biased were located in the surrounding grasslands and forest, estimates (Link 2003, 2004); moreover, the derived especially along the slopes and valleys on either side of CIs are likely to underestimate sampling uncertainty 1 main ridge in the central part of the study area. If because the data are too sparse to assess actual the bears were present in that area, their radio signals heterogeneity. Certainly, with small numbers of would have been clearly heard and their locations recaptures, the buffer size used to estimate density easily obtained by triangulation. When bears traveled

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 53 beyond this area, however, they were more difficult to locate by triangulation, so often just a single radio bearing (or multiple non-intersecting bearings) was obtained. Other times, when the bears traveled even farther beyond the central study area, their radio signals could not be heard at all. After a full year of radiotracking, 1 bear was located in the central study area on 64% of tracking attempts and the other located on 70% of tracking attempts. The portions of their home ranges within this area highly overlapped (Fig. 2). Circumscribing the combined area of use yields an area of ,12 km2. Hence, during the course of a year, 1 of the bears was present 64% of the time and the other 70% of the time within this 12-km2 area; this yields a mean of 0.64 + 0.70 5 1.34 bear-equivalents/12 km2,or11 bears/100 km2. Occasionally (although rarely), some other bears were observed in the area, so it is probably safe to assume that 12 bears/100 km2 is a minimum density estimate. Importantly, this repre- sents the average density over the course of a year — at times there might have been 3 or 4 bears present [25–33/100km2], either by chance or due to avail- ability of some type of food, and at other times 0 or 1 [0–8/100km2]). Although this is not a formal estimate of density, because it does not rigorously account for bears other than the 2 that were collared, it certainly Fig. 2. Portions of home ranges of 2 Andean bears in northern Bolivia (adapted from Paisley 2001). represents a minimum and covers a much longer Convex polygons enclose locations obtained by period than the 1-month camera-trapping study in triangulation over a year, and outer rectangle the same area. Notably, this minimum density is 2– (12 km2) circumscribes both polygons. The 2 bears 3x higher than that reported in the camera-trapping were located within this core area 64% and 70% of the time that they were radiotracked, but were study; however, it is very close to the estimate of 13.6 2 outside this area the rest of the time (i.e., either their bears/100 km that was rejected in the camera- radio signal could not be heard or the signal came trapping study because the buffer was deemed to be from outside this area, but location could not be unreasonably small (Rı´os-Uzeda et al. 2007). Iron- determined). These data yield a density estimate ically, the more reasonable camera-trapping buffer, higher than that obtained by camera trapping in the same area 5 years later (Fig. 1). taken directly from Paisley’s (2001) telemetry data, yielded a much lower density estimate than derived 3. The camera-trapping study site included directly from Paisley’s data. areas, possibly of lower density, outside the Several possible explanations exist for the discrep- 12-km2 telemetry study site. ancy between these 2 density estimates. 4. The density estimate from the camera-trap- ping study was biased low due to 1 or more 1. Density declined during the 5-year interlude of the factors described in the previous between the studies (1999–2004). section. 2. Density in this high-altitude area was lower 5. The estimates are a function of small study during the height of the dry season (Aug– areas and small sample sizes and are not Sep), when the camera-trapping study was really biologically different. conducted, than during other times of year because the bears tended to use other Any or all of these may have contributed, but I areas. favor point (5) as the primary explanation.

Ursus 22(1):47–64 (2011) 54 ANDEAN BEAR POPULATION ESTIMATES N Garshelis

Local population estimate based on DNA hair captures Undoubtedly the most rigorous estimate of abundance and density of Andean bears yet obtained was that of Viteri (2007). Her study was a mark– recapture design where the marks were DNA identifications of hair samples (Woods et al. 1999, Boulanger et al. 2004a, Lukacs and Burnham 2005a). Hair samples were collected by stringing 2 strands of barbed wire around tree trunks and wooden posts to form a small enclosure and putting a scent lure in the middle. Bears attracted to the lure crawled over or under the wires, and in so doing, left some hair with follicles (containing DNA) on the barbs. This study was conducted within the Cayambe- Coca Ecological Reserve in northern Ecuador. Hair 2 2 Fig. 3. Grid cells (3 x 3km , n = 94) in the Cayambe- traps were set out in a grid pattern of 9-km cells Coca Ecological Reserve, Ecuador, where Andean (3 km apart). This cell size was chosen based on bears were hair-trapped to produce a population presumed movements of bears to afford each bear estimate (adapted from Viteri 2007). Study area cells living in the area an opportunity to encounter at with no hair trap were omitted (except 1 interior cell marked with X). Quality of habitat in each cell was least 1 trap. A total of 103 grid cells (94 of which had categorized as prime (class 1), moderate (class 2–3), 2 a hair trap) encompassed 927 km , a reasonably or poor (class 4–5), based on classes of habitat large study area size. Sampling was conducted over a suitability (Cuesta et al. 2003), derived by modeling period of .3 months, during which time each cell incidence of bear sign in this same area. I catego- rized cells as to the predominant habitat suitability was checked and hairs collected 4 times. class by visually overlaying Cuesta et al.’s (2003:Fig. Viteri (2007) obtained genotypic information for 3) habitat suitability map with some extrapolations to 54 samples, from which she identified 25 individual cells outside their study area boundaries using bears. Various mark–recapture estimators (Pro- Viteri’s (2007) habitat map. Individual identifications grams CAPTURE and CAPWIRE; Otis et al. and gender of hair-trapped bears were determined by genetic analysis (Viteri 2007). The number of 1978, Miller et al. 2005) yielded population estimates different individual bears and their sex (M = male, F = of 33–48 bears (95% CI for full range of estimates: female, U = unknown) identified in the eastern, 26–69). Converting these estimates to density re- center, and western portions of the study site is quires population closure; accordingly, this study shown above the map. These 3 geographic areas were delineated based on predominant habitat, for was established within boundaries that were thought illustrative purposes. Captures in all areas were to be barriers to bear movements (i.e., high ridges male-dominated, and capture rate did not corre- and presumed marginal habitat). An analytical test spond with habitat suitability: twice as many bears indicated that the population was closed, but this were captured in the west, where habitat for bears pertains only to demographic, not geographic was considered mainly poor, as in the east, where much of the habitat was considered prime, with closure. approximately the same trapping effort in each (n = A highly-skewed sex ratio in all portions of the 28 and 26 hair-traps respectively). Some bears were study area (Fig. 3) suggests that the study area was detected in .1 portion of the study site, so totaling not geographically closed: in total, 15 (65%) of the the number captured in each geographic area (31) exceeds the total number of different bears identified identified bears were males and only 8 were females across the whole area (25). (2 were unidentified as to sex). Sex ratios of captures could be skewed by a combination of the following: (1) an uneven sex ratio in the living population, (2) protected area, but even with some exploitation by unequal vulnerability to capture, or (3) greater humans, the sex ratio of the living population would, movements by the more commonly captured sex if anything, be expected to become more female- (males in this case). This population was within a biased because male bears are more prone to being

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 55 killed when in conflict situations (Goldstein et al. were captured in what had been categorized as prime 2006, Oi 2009). I suspect the capture of more males habitat (Fig. 3). Indeed, 35% of bear sign in Cuesta than females partly reflected the larger home ranges et al.’s (2003) study within the same area was found of males: the study area would be expected to in montane habitat (which comprised encompass a portion of more male than female home about 5% of the study area), but only ,12% of ranges, including some males that lived mainly Viteri’s (2007) captures of identifiable bears occurred outside its borders. in this habitat (my calculation); standardized for Skewed sex ratios in capture samples, suggesting season (May–Aug hair collection), .40% of bear lack of geographic closure for wide-ranging males, sign was found in this habitat. Conversely, the least are common in studies of bears (Boulanger and suitable habitat was judged to be a high-elevation McLellan 2001, Boulanger et al. 2004a, Mowat et al. (.3,600 m) section of grassland pa´ramo in the west 2005). Combining the sexes can severely bias the that was traversed by a road, and a low elevation resulting estimates of abundance and density (Harm- (,1,900 m) section in the east with a road and foot sen et al. 2011). One way to circumvent this is to path connecting 2 towns (Cuesta et al. 2003). restrict the estimate to the sex that exhibits sufficient However, as many as 7 (28%) different bears were closure (Harris et al. 2010). captured in this anthropogenically-disturbed habitat Assuming geographic closure and simply dividing that was perceived, from sign, to be rarely used by the lowest and highest population estimates within bears. This mismatch between the distribution of the 95% CI by the size of the study area yields a sign and hair captures is difficult to reconcile, but range of 3–7 bears/100 km2 for Viteri’s (2007) study. could suggest an under-sampling of bears in the If the study area were not closed, allowing bears to prime habitat (and thus underestimation of density). wander in and out, then the effective sampling area Possibly the scent of the lure did not carry as far in (as discussed previously with regard to camera the forested habitat. Alternately, the detection of trapping) was larger and the actual density lower sign was biased. However, sign was commonly than this simple estimate. However, even without a detected in the grassland pa´ramo at sites distant correction for lack of closure, these density estimates from roads (mean incidence 5 28% of all sign), and are already quite low, and moreover, they probably scats and feeding signs in pa´ramo tend to last include some cubs (because 2 strands of barbed wire considerably longer than in cloud forest (Paisley were used to collect the hair samples). What is 2001). This suggests that sign surveys would unclear, though, is whether some bears, notably overestimate (not underestimate) bear use of females, might not have been attracted to the lure or pa´ramo, making the apparent high rate of hair were behaviorally averse to crawling through the captures in this habitat even more enigmatic. A barbed wire (Woods et al. 1999), making them potential explanation is that DNA in hair samples essentially uncatchable, at least over the short term from cloud forest habitat degraded more quickly (M. Gibeau [Parks Canada, Lake Louise, Alberta, than in samples from the pa´ramo. Thus, whereas 54 Canada, personal communication, 2010] observed samples provided identifications of 25 different this for grizzly bears filmed with remote video bears, another 94 samples yielded insufficient genetic cameras). If heterogeneity in capture probabilities information to be included in the analysis (Viteri was the primary explanation for the skewed sex 2007), and these may have occurred more often in ratio, then the population and corresponding density the forested habitat. estimates would have been biased low despite the lack of geographic closure. The population estima- tors used in this study were designed to correct for Bias and precision of estimates capture heterogeneity, but cannot account for The accuracy and precision of a population animals with very low capture rates (Link 2003, estimate are largely a function of sampling. One 2004). can never expect to achieve a perfectly accurate A particularly enigmatic finding was that a high estimate, but with increased sampling, the estimated proportion of the bears were captured in portions of value should approach the true value, and certainty the study area that were previously identified as poor in the estimate should increase. Indeed, precision Andean bear habitat, based on incidence of bear sign typically increases (CIs become narrower) with (Cuesta et al. 2003), and a relatively low proportion larger samples. However, CIs only account for

Ursus 22(1):47–64 (2011) 56 ANDEAN BEAR POPULATION ESTIMATES N Garshelis variation within the sample data; they provide no has been used as an indicator of relative habitat indication of bias. quality for other species of bears (Apps et al. 2004, Robson and Regier (1964) proposed that for Nams et al. 2006, Proctor et al. 2010). Only if there is management purposes (which would include conser- a mismatch between the spatial distribution of vation monitoring), investigators should strive to captures and some other independent assessment of ensure that there is a 95% probability that popula- habitat quality (e.g., Viteri 2007 versus Cuesta et al. tion estimates are within 25% of true population size. 2003) would there be cause for concern. This desired precision suggests that upper and lower The underlying cause of a skewed sex ratio might limits of the 95% CI should be within 25% of the be revealed by comparing recapture rates of males point estimate. Assuming that the sampling is not and females (Kendall et al. 2009). In the extreme biased, one can calculate necessary sample sizes to case, where the capture sex ratio is very lopsided achieve this. In the simplest form of mark–recapture apparently due to 1 sex being under-sampled (e.g., population estimation (2-sample Petersen estimate), females avoiding hair traps) or over-sampled (e.g., a recapture sample that includes 16 marked animals males routinely entering and leaving the study area), would tend to produce 95% confidence limits within a population estimate closer to truth might be 50% of the point estimate; a recapture sample with derived by selecting the sex with the presumably ,64 marked animals would be required to achieve better estimate and doubling it (Bellemain et al. the desired 25% (C.J. Schwarz, 2006, An introduc- 2005, Peacock et al. 2011). Whereas such an tion to mark–recapture methods used in fisheries approach may not seem analytically rigorous, it management, unpublished report, Simon Fraser must be stressed that analytical rigor typically University, Burnaby, British Columbia, Canada). cannot overcome sampling biases, and extreme This sort of sample size may be impossible to attain efforts to correct such biases may be necessary. for Andean bears with some techniques (e.g., camera A precise but biased estimate may be especially trapping), although the same level of precision could misleading because the true abundance (or density) be achieved with smaller sample sizes within multiple could be well outside the CI. Sharma et al. (2010) recapture sessions. demonstrated this with a camera-trapping study in a In attempting to obtain unbiased and precise reserve with a known number of (Panthera estimates, both the size and representativeness of the tigris, n 5 14). By excluding subsets of the accumu- sample are key. With very small samples, the lated photos (from specific cameras), they simulated likelihood increases that the data are biased by an the effects of varying camera trap densities on the odd event or unusual individual. In Paisley’s (2001) reliability of the population estimates. In a low study, for example, 1 of 2 radiocollared bears was a density tiger population with a low density of camera conspicuously under-sized adult and the other was a traps, only 34% of simulated estimates contained the subadult, so their movements and habitat use may true population size within their 95% CIs. not have represented the population at large (Paisley Camera trap-based population estimates are and Garshelis 2006). Thus, although the density particularly difficult to evaluate for bias because estimate derived from the home range use of these gender information is usually lacking and, in some bears accurately portrayed the bear density within cases at least, a significant number of photos may be the small area in which they spent most of their time, unidentifiable to individual because of inadequate that density may not have reflected bear density in clarity, framing, or position of the bear (Zug 2009). similar habitats across that region of Bolivia. In this Also, bears with similar markings could be misclas- study, the bears were captured and handled, and sified. A method to quantify potential errors in their weights, measurements, and body conditions identification and incorporate them into the CIs has helped in assessing whether they were normal. not yet been developed, but possibly could be Camera trap photos and DNA do not enable such patterned after methods devised for genotyping an assessment. errors (Lukacs and Burnham 2005b, Petit and For DNA-based studies, sex ratio and spatial Valiere 2006, Knapp et al. 2009, Wright et al. distribution of the capture samples provide the 2009). However, if these sorts of events are non- primary indications of representativeness. Spatial random (e.g., if some similar-looking individuals are distribution of captures is expected to vary in a non- often confused), or even if they are random but uniform environment. In fact, spatial distribution capture success is extremely low, this sort of

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 57 incomplete, heterogeneous sampling can significant- process that helps to correct mistakes and misinter- ly bias the results (Link 2003, 2004). Even with pretations and provide better direction for the reasonably large capture samples, heterogeneous future. This process is especially important for sampling is common in bear studies, often resulting threatened bear populations because many of the in significant underestimates of population size techniques regularly employed to track changes in (Noyce et al. 2001). their population sizes have produced misleading or To ensure reliable results from mark–recapture uncertain results (Garshelis 2002). An unfortunate models, average capture probabilities should gener- irony is that the poorest data are available for the ally be .0.2 (White et al. 1982, Boulanger et al. species and populations of bears that are most 2004a). For small populations (e.g., ,50 animals), threatened. they should be higher, to account for heterogeneity, Certainly there are ample reasons to question the avoid bias, and compute meaningful CIs. High trap reliability of all of these estimates for Andean bears. densities are needed to achieve such high capture Peyton et al.’s (1998) estimates, based on extrapola- probabilities if home ranges are fairly small (Settlage tion of American black bear densities, span a wide et al. 2008). Capture rates are bound to be low for range (18,000–65,000 bears) and so are much less low-density Andean bear populations living in areas precise than indicated by conservative reports of just where it is logistically difficult to set and check a the lower value, or a rounded-off value of ,20,000. large number of camera traps over a wide enough Some have nevertheless used the lowest density area to sample many individuals. Rı´os-Uzeda et al. figure to derive separate range–country estimates. (2007) calculated a probability of capture of only Obviously, no real assessment of reliability is 0.08 for their camera-trapping study and estimated a available for such estimates because they were based population size of only 3 bears. Biologists thus face on a surrogate species. Peyton et al.’s (1998) use of the dilemma of whether employing an analytical the black bear density estimates filled an information approach that is poorly equipped to handle such void and was arguably somewhat better than a small samples returns results that are reliable enough guess. Others have produced abundance and density to be of use in conserving species like the Andean guesstimates for other species of bears based only on bear. If capture rates are low and sample sizes small, occasional sightings and subjective assessments of investigators are unlikely to discern individual amount of sign that have been extrapolated over capture heterogeneity (which is nearly ubiquitous) large areas (e.g., sun bears [Helarctos malayanus], and therefore would be unaware of the negative bias Davies and Payne 1981; sloth bears [Melursus in the resulting population estimate (Ebert et al. ursinus], Yoganand et al. 2006). 2010, Proctor et al. 2010, Harmsen et al. 2011). Ruiz-Garcia’s (2003) genetic diversity-based esti- Whereas such underestimation errs on the side of mates also span a wide range. However, this range demanding more conservation attention to the was trimmed through a process that was not easily population, it also may hamper the ability to defensible or replicable. The apparent concordance recognize conservation efforts that have succeeded between the estimates of Ruiz-Garcia’s (2003) and in fostering population growth. Peyton et al. (1998) owe in part to an effort to frame the genetics estimates by what was considered to be field census data. The central focus of Ruiz-Garcia’s Reliability of Andean bear estimates study was genetic heterogeneity and gene flow, but Reliability of population estimates can be gauged again, in an attempt to fill a void, the methods may from their precision, potential biases, and transpar- have been stretched beyond their level of reliability ency of methods, results, and analyses (Conroy and to generate population estimates. Carroll 2009). My review was intended to critically Rı´os-Uzeda et al. (2007) specifically sought to evaluate these aspects of reliability for Andean bear estimate density using camera-trapping, a technique estimates. The issues raised here were not intended to that had been successfully employed on other disparage the work of the investigators who strove, species, mainly with distinctive pelages. Indeed, sometimes with enormous effort, to provide infor- some of the same investigators had previously used mation to the scientific and conservation communi- camera traps to estimate abundance and density of ties that could aid in understanding and protecting jaguars (Wallace et al. 2003) in one of the same populations of these bears. Critical evaluation is a protected areas (Madidi National Park) where the

Ursus 22(1):47–64 (2011) 58 ANDEAN BEAR POPULATION ESTIMATES N Garshelis bears were later studied. The study also sampling was random, which it was not, given yielded a small number of captures and identifiable difficulties of sampling during the rainy season). This individuals (but twice that of the bear study). Using suggests a range of 1.22 (2 x 0.61) to 1.46 bear- only telemetry-based buffers to estimate effective equivalents in the 12-km2 area, or a density of 10.2– trapping area and calculate density, Rı´os-Uzeda et 12.2 bears/100 km2, which is probably far more al.’s (2007) 95% CI (2.6–8.5/100 km2) was rather precise than warranted given the nature of this crude imprecise; nevertheless, it suggested a low density, calculation. But, especially with the addition of other equivalent to or less than the lowest reported bears that were occasionally seen in this area, it American black bear density, used by Peyton et al. appears that this estimate is truly higher than the (1998). The issue, though, is whether the reported camera-trapping estimate in the same area. When 2 95% CI actually contained the true population estimates in the same area, based on different density in this area, or if it was significantly biased. methodologies, produce results with non-overlap- This problem is not unique to this study or to bears, ping CIs, the reliability of 1 or both estimates must but is likely pervasive for camera-trapping studies of be called into question. low-density species (Harmsen et al. 2011). Viteri’s (2007) estimate of Andean bear abundance It is important to note that the reported CIs for was well-designed and carefully conducted, yet the abundance estimates, when converted to density, precision of the estimate was low (within 30–50% of typically do not include the error associated with the point estimates) for monitoring purposes. Wide CIs, estimation of effective trapping area. So in the case though, are not uncommon for even the most of Rı´os-Uzeda et al.’s (2007) study, the calculated rigorous population estimates of bears (Miller et al. CIs assume that the effective area was a fixed, known 1997, Boulanger et al. 2002). Moreover, population quantity (or, rather a choice among 3 quantities, trends may be gleaned from a series of estimates with Fig. 1). Clearly that was not the case because the overlapping CIs (Garshelis and Hristienko 2006, buffer area was derived from estimates of home Clark et al. 2010), and estimates of population range sizes, which are likely to vary substantially by growth (l) can even be obtained from mark– individual, sex, method of calculation, and number recapture data that fail to produce a useful estimate of telemetry locations. Hence, the denominator in of abundance (e.g., Pradel model in Program the density calculation (area) probably has more MARK; Boulanger et al. 2004b, Proctor et al. 2010). uncertainty than the numerator (abundance), signif- Notably, Viteri’s (2007) estimates, converted to icantly reducing the reliability of the estimate. densities, were equivalent to those of Rı´os-Uzeda et The density estimate based on Paisley’s (2001) al. (2007) and would have been even lower if home range data was presented here with no corrected for effective trapping area (lack of closure). measure of precision. Certainly the area used by An important question, though, is whether these the 2 bears (Fig. 2) was not measured without error: results were actually biased low, given their discor- potential sources of error include telemetry triangu- dance with models of habitat quality judged from lation, sampling (even when the bears were within abundance of sign (Cuesta et al. 2003). Again, this general area, they could have been in other conflicting results reduce confidence in reliability. places when nobody was observing), and method of The DNA results, however, are at least highly depicting this portion of their home range (minimum transparent. Geneticists routinely report error rates convex polygon). However, these sources of error are associated with genotyping, and methods exist for essentially irrelevant because the area of use was assessing and correcting effects of these errors on taken to be a rectangle encompassing both bears, a population estimates (Lukacs and Burnham 2005b, very crude approximation that would tend to Petit and Valiere 2006, Knapp et al. 2009, Wright et underestimate density (because it includes areas al. 2009). Thus, although genetic identifications where the bears were never detected; Fig. 2). Hence outwardly seem less apparent than photographs, the only remaining major source of error would be they are actually more strictly defined, due to the proportion of time that the bears spent in this standardized and validated laboratory procedures area. Combining data for the 2 bears (because they (Waits and Paetkau 2005), and are thus repeatable spent a similar proportion of time in the area), a and defensible. Unfortunately, individual identifica- simple CI can be calculated (146 locations in 218 tions from camera trap photos are generally not attempts 5 0.67, 95% CI 5 0.61–0.73, presuming the subject to the same strict standards so might be

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 59 judged to be less reliable. That is not to say that less The existing population estimates for Andean care is taken in matching photographs. Indeed, other bears, reviewed and evaluated here, are all useful in studies have reported exacting protocols (Jackson et one respect. They are a good demonstration of a al. 2006, Larrucea et al. 2007), and even computer- variety of approaches and highlight the kinds of aided photographic matching, which is helpful (or biases and other shortcomings that are likely to be even necessary) for complex pelage patterns (Kelly encountered in future studies. They exemplify the 2001, Speed et al. 2007, Hastings et al. 2008, Sherley desire to better understand Andean bear populations et al. 2010). In terms of Andean bears, however, as well as the difficulties of doing so (specifically, pattern matching often relies on just a few white issues of small sample size and sampling biases). facial streaks, which can appear slightly different at Others studying small bear populations have noted different angles, different lighting, or different the benefit of combining different sorts of sampling wetness, and different observers may employ differ- — for example, setting some hair snares at feeding ent personal thresholds for distinguishing individu- sites where densities are relatively high, plucking als. Some efforts have recently been made to better hairs off trees, and obtaining DNA from scats. define and standardize the process used to differen- Doing so not only adds to the sample size, but also tiate individuals of this species from camera-trap helps to expose more of the population heterogeneity photographs (Jones 2010). Improvements to this (Boulanger et al. 2008, Gervasi et al. 2008, De Barba methodology will be crucial if camera-trapping et al. 2010, Stetz et al. 2010). If a bear eludes capture studies are to be used to derive reliable estimates of through 1 sampling procedure, it might be captured abundance (and survival rates), which will require another way. Additionally, more reliable population larger samples and greater certainty in recapture estimates may accrue from applying multiple ana- events than obtained thus far. lytical techniques, especially where sample size is limited by small population size (Meijer et al. 2008). Collectively, results obtained so far suggest that Usefulness of population estimates Andean bears may tend to occur at lower densities It seems almost tautological that if a population or than most hunted populations of American black density estimate is reliable, it would be useful, but is bears (Garshelis 1994); in this respect, they may be there any value in less reliable estimates? Potentially, more akin to high-density interior populations of a collection of less rigorous estimates, all in the same brown bears (Schwartz et al. 2003). If this is upheld ballpark, could assist conservation assessments (e.g., through further work, it would indicate that these evaluating risks and effects of ; Sa´nchez- bears have a relatively low resilience to human Mercado et al. 2008). Conversely, highly-varying exploitation. estimates might be useful for identifying relative Effective conservation strategies for Andean bears population strongholds and thereby defining con- have been developed in the absence of reliable servation priority areas. Previously, such areas have population estimates (Rodrı´guez et al. 2003, Castel- been identified based on incidence of sign (Peralvo et lanos et al. 2010). To that extent, it could be argued al. 2005), but sign deposition and detection can vary that accurate estimates are not really necessary, by habitat type, so difficulties may arise in compar- especially for a species that is not managed through ing areas with different habitat compositions. harvest. On the other hand, forecasts for the Depending on the particular circumstances, this sort persistence of this species within the many small, of bias may either be diminished or exacerbated by isolated patches of habitat in which it tends to exist attempting to progress from an index to an estimate. depend largely on how many bears there really are For example, Rı´os-Uzeda et al. (2007) found it (Kattan et al. 2004). difficult to obtain camera captures of bears in high The studies reviewed here suggest that these bears altitude grasslands because vegetation caused innu- may occur at much lower densities than once merable false camera-triggers, whereas sign was thought, a finding, which if corroborated through readily observable in this habitat (Rı´os-Uzeda et al. further research, has strong implications for conser- 2005). Obviously, researchers and conservationists vation planning (e.g., assessing the value of protect- must decide when the extra effort of obtaining a ing habitat fragments of a certain size or the need to population estimate is warranted, and what level of link small areas; Yerena et al. 2003, Kattan et al. accuracy is required. 2004). However, because of the difficulty of con-

Ursus 22(1):47–64 (2011) 60 ANDEAN BEAR POPULATION ESTIMATES N Garshelis

ducting studies that are likely to return reliable ———, G.C. WHITE, B.N. MCLELLAN,J.WOODS,M. abundance estimates, such work should not supplant PROCTOR, AND S. HIMMER. 2002. A meta-analysis of the many other ecological, behavioral, and demo- grizzly bear DNA mark–recapture projects in British graphic studies that are needed to better understand Columbia, Canada. Ursus 13:137–152. this species. Only through careful attention to the ———, S. HIMMER, AND C. SWAN. 2004b. Monitoring of selection of a study site, study design, representative grizzly bear population trends and demography using DNA mark–recapture methods in the Owikeno Lake sampling, analytical approach, and critical scrutiny area of British Columbia. Canadian Journal of Zoology of the results will population estimates be reasonably 82:1267–1277. reliable, but even with concern for all these details, ———, ———, J.G. WOODS, M.F. PROCTOR, AND C. the many difficult circumstances associated with STROBECK. 2004a. Sampling design and bias in DNA- studying this species will make it much harder to based capture–mark–recapture population and density obtain estimates with the same level of reliability as estimates of grizzly bears. Journal of Wildlife Manage- achieved for North American bears. Consequently, ment 68:457–469. efforts directed toward population monitoring meth- ———, K.C. KENDALL, J.B. STETZ, D.A. ROON, L.P. ods that do not require reliable population estimates WAITS, AND D. PAETKAU. 2008. Multiple data sources may be more useful. The same could probably be improve DNA-based mark–recapture population esti- said for most populations of bears in Asia; mates of grizzly bears. Ecological Applications accordingly, this review may be helpful to bear 18:577–589. biologists in that region as well. CASTELLANOS, A. 2004. Andean bear research in the Intag Region, Ecuador. International Bear News 13(2):25–26. ———, J. CEVALLOS,A.LAGUNA,L.ACHIG,P.VITERI, AND S. MOLINA. 2010. National strategy for Andean bear Acknowledgments conservation. Ministry of Environment of Ecuador, I am especially thankful to I. Goldstein for Anyma Printing, , Ecuador. (In Spanish.) inviting me to present this paper at the Second ———. 2011. Andean bear home ranges in the Intag International Symposium on the Andean Bear, held Region, Ecuador. Ursus 22:65–73. in Lima, Peru, November 2008. Symposium spon- CLARK, J.D., R. EASTRIDGE, AND M.J. HOOKER. 2010. sors provided financial support. R. Wallace, M. P. Effects of exploitation on black bear populations at Viteri, and 3 anonymous reviewers provided helpful White River National Wildlife Refuge. Journal of comments on an earlier version of this manuscript. Wildlife Management 74:1448–1456. CONROY, M.J., AND J.P. CARROLL. 2009. Quantitative conservation of vertebrates. Wiley-Blackwell, Chiche- Literature cited ster, UK. UESTA ERALVO AND VAN ANEN APPS, C.D., B.N. MCLELLAN, J.G. WOODS, AND M.F. C , F., M.F. P , F.T. M . 2003. PROCTOR. 2004. Estimating grizzly bear distribution and Andean bear habitat use in the Oyacachi River Basin, abundance relative to habitat and human influence. Ecuador. Ursus 14:198–209. Journal of Wildlife Management 68:138–152. DAVIES, G., AND J. PAYNE. 1981. A faunal survey of Sabah. ASPI, J., E. ROININEN,M.RUOKONEN,I.KOJOLA, AND C. World Wildlife Fund Malaysia, Project 1692, Kuala VILA` . 2006. Genetic diversity, population structure, Lampur, Malaysia. effective population size and demographic history of DE BARBA, M., L.P. WAITS,P.GENOVESI,E.RANDI,R. the Finnish population. Molecular Ecology CHIRICHELLA, AND E. CETTO. 2010. Comparing oppor- 15:1561–1576. tunistic and systematic sampling methods for non- BALME, G.A., L.T.B. HUNTER, AND R. SLOTOW. 2009. invasive genetic monitoring of a small translocated Evaluating methods for counting cryptic carnivores. brown bear population. Journal of Applied Ecology Journal of Wildlife Management 73:433–441. 47:172–181. BELLEMAIN, E., J.E. SWENSON,D.TALLMON,S.BRUNBERG, DILLON, A., AND M.J. KELLY. 2008. home range, AND P. TABERLET. 2005. Estimating population size of overlap and density: Comparing radio telemetry with elusive animals with DNA from hunter-collected feces: camera trapping. Journal of Zoology 275:391–398. Four methods for brown bears. Conservation Biology DREHER, B.P., S.R. WINTERSTEIN, K.M. SCRIBNER, P.M. 19:150–161. LUKACS, D.R. ETTER, G.J.M. ROSA, V.A. LOPEZ,S. BOULANGER, J., AND B.N. MCLELLAN. 2001. Closure LIBANTS, AND K.B. FILCEK. 2007. Noninvasive estima- violation in DNA-based mark–recapture estimation of tion of black bear abundance incorporating genotyping grizzly bear populations. Canadian Journal of Zoology errors and harvested bear. Journal of Wildlife Man- 79:642–651. agement 71:2684–2693.

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 61

EBERT, C., F. KNAUER,I.STORCH, AND U. HOHMANN. 2010. of relative abundance and density in the Individual heterogeneity as a pitfall in population eastern rainforests of Madagascar using remotely- estimates based on non-invasive genetic sampling: a triggered camera traps. Oryx 44:219–222. review and recommendations. Wildlife Biology 16:225– GERVASI, V., P. CIUCCI,J.BOULANGER,M.POSILLICO,C. 240. SULLI,S.FOCARDI,E.RANDI, AND L. BOITANI. 2008. A EFFORD, M.G. 2004a. Density estimation in live-trapping preliminary estimate of the Apennine brown bear studies. Oikos 106:598–610. population size based on hair-snag sampling and ———, D.K. DAWSON, AND C.S. ROBBINS. 2004b. DEN- multiple data source mark–recapture Huggins models. SITY: Software for analyzing capture–recapture data Ursus 19:105–121. from passive detector arrays. Animal Biodiversity and GOLDSTEIN, I., S. PAISLEY,R.WALLACE, J.P. JORGENSON,F. Conservation 27:217–228. CUESTA, AND A. CASTELLANOS. 2006. Andean bear– ———, D.L. BORCHERS, AND A.E. BYROM. 2009. Density livestock conflicts: A review. Ursus 17:8–15. estimation by spatially explicit capture–recapture: GOOSSENS, B., L. CHIKHI,M.ANCRENAZ,I.LACKMAN- Likelihood-based methods. Pages 255–269 in D.L. ANCRENAZ,P.ANDAU, AND M. BRUFORD. 2006. Genetic Thompson, E.G. Cooch, and M.J. Conroy, editors. signature of anthropogenic population collapse in Modeling demographic processes in marked popula- orang-utans. PLoS Biology 4:285–291. tions. Springer, New York, New York, USA. HARMSEN, B.J., R.J. FOSTER, AND C.P. DONCASTER. 2011. ELKAN, P.W. 2003. Ecology and conservation of bongo Heterogeneous capture rates in low density populations antelope (Tragelaphus eurycerus) in lowland forest, and consequences for capture–recapture analysis of northern Republic of Congo. Dissertation, University camera-trap data. Population Ecology 53:253–259. of Minnesota, Minneapolis, Minnesota, USA. HARRIS, R.B., AND F.W. ALLENDORF. 1989. Genetically FRANKHAM, R. 1995. Effective population size/adult effective population size of large : An assess- population size ratios in wildlife: a review. Genetical ment of estimators. Conservation Biology 3:181–191. Research 66:95–107. ———, J. WINNIE,JR., S.J. AMISH,A.BEJA-PEREIRA,R. ———. 1996. Relationship of genetic variation to popu- GODINHO,V.COSTA, AND G. LUIKART. 2010. Argali lation size in wildlife. Conservation Biology 10:1500– abundance in the Afghan Pamir using capture–recap- 1508. ture modeling from fecal DNA. Journal of Wildlife GARDNER, B., J.A. ROYLE, AND M.T. WEGAN. 2009. Management 74:668–677. Hierarchical models for estimating density from DNA HASTINGS, K.K., L.A. HIBY, AND R.J. SMALL. 2008. mark–recapture studies. Ecology 90:1106–1115. Evaluation of a computer-assisted photograph-match- ———, ———, ———, R. RAINBOLT, AND P.D. CURTIS. ing system to monitor naturally marked harbor seals at 2010. Estimating black bear density using DNA data Tugidak Island, Alaska. Journal of Mammalogy from hair snares. Journal of Wildlife Management 89:1201–1211. 74:318–325. IMMELL, D., AND R.G. ANTHONY. 2008. Estimation of black GARSHELIS, D.L. 1992. Mark–recapture density estimation bear abundance using a discrete DNA sampling device. for animals with large home ranges. Pages 1098–1111 in Journal of Wildlife Management 72:324–330. D.R. McCullough and R.H. Barrett, editors. Wildlife JACKSON, R.M., J.D. ROE,R.WANGCHUCK, AND D.O. 2001: Populations. Elsevier Applied Science, London, HUNTER. 2006. Estimating snow population UK. abundance using photography and capture–recapture ———. 1994. Density-dependent population regulation of techniques. Wildlife Society Bulletin 34:772–781. black bears. Pages 3–14 in M. Taylor, editor. Density- JONES, T. 2010. Detection probability and individual dependent population regulation in black, brown, and identification of the Andean bear (Tremarctos ornatus) polar bears. International Conference on Bear Research using camera trapping methods. Thesis, University of and Management Monograph Series 3. Wisconsin, Madison, Wisconsin, USA. ———. 2002. Misconceptions, ironies, and uncertainties JORGENSON, J.P., AND S. SANDOVAL-A. 2005. Andean bear regarding trends in bear populations. Ursus 13:321–334. management needs and interactions with humans in ———, AND H. HRISTIENKO. 2006. State and provincial Colombia. Ursus 16:108–116. estimates of American black bear numbers versus KARANTH, K.U. 1995. Estimating tiger Panthera tigris assessments of population trend. Ursus 17:1–7. populations from camera-trap data using capture— ———, AND K.V. NOYCE. 2006. Discerning biases in a recapture models. Biological Conservation 71:333–338. large scale mark–recapture population estimate for ———, AND J.D. NICHOLS. 1998. Estimation of tiger black bears. Journal of Wildlife Management densities in using photographic captures and 70:1634–1643. recaptures. Ecology 79:2852–2862. GERBER, B., S.M. KARPANTY,C.CRAWFORD,M.KOTSCH- KATTAN, G., O.L. HERNA´ NDEZ,I.GOLDSTEIN,V.ROJAS,O. WAR, AND J. RANDRIANANTENAINA. 2010. An assessment MURILLO,C.GO´ MEZ,H.RESTREPO, AND F. CUESTA.

Ursus 22(1):47–64 (2011) 62 ANDEAN BEAR POPULATION ESTIMATES N Garshelis

2004. Range fragmentation in the MEIJER, T., K. NORE´ N,P.HELLSTRO¨ M,L.DALE´ N, AND A. Tremarctos ornatus in the northern . Oryx ANGERBJO¨ RN. 2008. Estimating population parameters 38:155–163. in a threatened arctic population using molecular KELLY, M. 2001. Computer-aided photographic matching tracking and traditional field methods. Animal Con- in studies using individual identification: An example servation 11:330–338. from Serengeti . Journal of Mammalogy MILLER, C.R., AND L.P. WAITS. 2003. The history of 82:440–449. effective population size and genetic diversity in the KENDALL, K., J. STETZ,J.BOULANGER, A.C. MAClEOD,D. Yellowstone grizzly (Ursus arctos): Implications for PAETKAU, AND G.C. WHITE. 2009. Demography and conservation. Proceedings of the National Academy of genetic structure of a recovering grizzly bear popula- Sciences (USA) 100:4334–4339. tion. Journal of Wildlife Management 73:3–17. ———, P. JOYCE, AND L.P. WAITS. 2005. A new method for KNAPP, S.M., B.A. CRAIG, AND L.P. WAITS. 2009. estimating the size of small populations from genetic Incorporating genotyping error into non-invasive mark–recapture data. Molecular Ecology 14:1991– DNA-based mark–recapture population estimates. 2005. Journal of Wildlife Management 73:598–604. MILLER, S.D., E.F. BECKER, AND W.B. BALLARD. 1987. LARRUCEA, E.S., G. SERRA, M.M. JAEGER, AND R.H. Black and brown bear density estimates using modified BARRETT. 2007. Censusing using remote cam- capture–recapture techniques in Alaska. International eras. Western North American Naturalist 67:538–548. Conference on Bear Research and Management LINK, W.A. 2003. Nonidentifiability of population size 7:23–35. from capture–recapture data with heterogeneous detec- ———. 1994. Black bear reproduction and cub survivor- tion probabilities. Biometrics 59:1123–1130. ship in south-central Alaska. International Conference ———. 2004. Individual heterogeneity and identifiability on Bear Research and Management 9(1):263–273. in capture–recapture models. Animal Biodiversity and ———, G.C. WHITE, R.A. SELLERS, H.V. REYNOLDS, J.W. Conservation 27:87–91. SCHOEN,K.TITUS, V.G. BARNES,JR., R.B. SMITH, R.R. LUIKART, G., N. RYMAN, D.A. TALLMON, M.K. SCHWARTZ, NELSON, W.B. BALLARD, AND C.C. SCHWARTZ. 1997. AND F.W. ALLENDORF. 2010. Estimation of census and Brown and black bear density estimation in Alaska effective population sizes: The increasing usefulness of using radiotelemetry and replicated mark–resight tech- DNA-based approaches. Conservation Genetics 11: niques. Wildlife Monographs 133. 355–373. MOWAT, G., D.C. HEARD, D.R. SEIP, K.G. POOLE,G. LUKACS, P.M., AND K.P. BURNHAM. 2005a. Review of STENHOUSE, AND D.W. PAETKAU. 2005. Grizzly Ursus capture–recapture methods applicable to noninva- arctos and black bear U. americanus densities in the sive genetic sampling. Molecular Ecology 14:3909– interior mountains of North America. Wildlife Biology 3919. 11:31–48. ———, AND ———. 2005b. Estimating population size NAMS, V.O., G. MOWAT, AND M.A. PANIAN. 2006. from DNA based closed capture–recapture data Determining the spatial scale for conservation purposes incorporating genotyping error. Journal of Wildlife — An example with grizzly bears. Biological Conser- Management 69:396–403. vation 128:109–119. LYNAM, A.J., A. RABINOWITZ,T.MYINT,M.MAUNG, K.T. NOYCE, K.V., D.L. GARSHELIS, AND P.L. COY. 2001. LATT, AND S.H.T. PO. 2009. Estimating abundance with Differential vulnerability of black bears to trap and sparse data: Tigers in northern Myanmar. Population camera sampling and resulting biases in mark–recap- Ecology 51:115–121. ture estimates. Ursus 12:211–226. MAFFEI, L., AND A.J. NOSS. 2008. How small is too small? OBBARD, M., AND E.J. HOWE. 2008. Demography of black Camera trap survey areas and density estimates for bears in hunted and unhunted areas of the boreal forest ocelots in the Bolivian Chaco. Biotropica 40:71–75. of Ontario. Journal of Wildlife Management 72:869– MARNEWICK, K., P.J. FUNSTON, AND K.U. KARANTH. 2008. 880. Evaluating camera trapping as a method for estimating ———, ———, AND C.J. KYLE. 2010. Empirical compar- abundance in ranching areas. South African ison of density estimators for large carnivores. Journal Journal of Wildlife Research 38:59–65. of Applied Ecology 47:76–84. MATTHEWS, S.M., R.T. GOLIGHTY, AND J.M. HIGLEY. 2008. OI, T. 2009. Anthropogenic mortality of Asiatic black Mark–resight density estimation for American black bears in two populations in northern Honshu, Japan. bears in Hoopa, California. Ursus 19:13–21. Ursus 20:22–29. MCCARTHY, K.P., T.K. FULLER,M.MING, T.M. OTIS, D.L., K.P. BURNHAM, G.C. WHITE, AND D.R. MCCARTHY,L.WAITS, AND K. JUMABAEV. 2008. ANDERSON. 1978. Statistical inference from capture Assessing estimators of abundance. data on closed animal populations. Wildlife Mono- Journal of Wildlife Management 72:1826–1833. graphs 62.

Ursus 22(1):47–64 (2011) ANDEAN BEAR POPULATION ESTIMATES N Garshelis 63

PAISLEY, S. 2001. Andean bears and people in Apolo- ROBSON, D.S., AND H.A. REGIER. 1964. Sample size in bamba, Bolivia: Culture, conflict and conservation. Petersen mark–recapture experiments. Transactions of Dissertation, University of Kent, Durrell Institute of the American Fisheries Society 93:215–226. Conservation and Ecology, Canterbury, UK. RODRI´GUEZ, D., F. CUESTA,I.GOLDSTEIN, L.G. NARANJO, ———, AND D.L. GARSHELIS. 2006. Activity patterns and O.L. HERNA´ NDEZ, AND A.E. BRACHO. 2003. Ecoregional time budgets of Andean bears (Tremarctos ornatus)in strategy for the conservation of the spectacled bear the Apolobamba Range of Bolivia. Journal of Zoology (Tremarctos ornatus) in the northern Andes. World 268:25–34. Wildlife Fund Colombia, Fundacio´n Wii, EcoCiencia, PARMENTER, R.R., T.L. YATES, D.R. ANDERSON, K.P. Wildlife Conservation Society. BURNHAM, J.L. DUNNUM, A.B. FRANKLIN, M.T. FRIG- ROYLE, J.A., J.D. NICHOLS, K.U. KARANTH, AND A.M. GENS, B.C. LUBOW,M.MILLER, G.S. OLSON, C.A. GOPALASWAMY. 2009. A hierarchical model for estimat- PARMENTER,J.POLLARD,E.REXTAD, T.M. SHENK, T.R. ing density in camera-trap studies. Journal of Applied STANLEY, AND G.C. WHITE. 2003. Small- Ecology 46:118–127. density estimation: A field comparison of grid-based RUIZ-GARCIA, M. 2003. Molecular population genetic vs. web-based density estimators. Ecological Mono- analysis of the spectacled bear (Tremarctos ornatus) graphs 73:1–26. in the northern Andean area. Hereditas 138:81–93. PEACOCK, E., K. TITUS, D.L. GARSHELIS, M.M. PEACOCK, ———, P. OROZCO-tERWENGEL,A.CASTELLANOS, AND L. AND M. KUC. 2011. Mark–recapture using tetracycline ARIAS. 2005. Microsatellite analysis of the spectacled and genetics reveal record-high bear density. Journal of bear (Tremarctos ornatus) across its range distribution. Wildlife Management 75:in press. Genes and Genetic Systems 80:57–69. PERALVO, M.F., F. CUESTA, AND F.T. VAN MANEN. 2005. SAARMA, U., S.Y.W. HO, O.G. PYBUS,M.KALJUSTE, I.L. Delineating priority habitat areas for the conservation TUMANOV,I.KOJOLA, A.A. VOROBIEV, N.I. MARKOV, of Andean bears in northern Ecuador. Ursus 16: A.P. SAVELJEV,H.VALDMANN, E.A. LYAPUNOVA, A.V. 222–233. ABRAMOV,P.MANNIL,M.KORSTEN,E.VULLA, S.V. PETIT, E., AND N. VALIERE. 2006. Estimating population PAZETNOV, V.S. PAZETNOV, S.V. PUTCHKOVSKIY, AND size with noninvasive capture–mark–recapture data. A.M. ROKOV. 2007. Mitogenetic structure of brown Conservation Biology 20:1062–1073. bears (Ursus arctos L.) in northeastern Europe and a PEYTON, B., E. YERENA, D.I. RUMIZ,J.JORGENSON, AND J. new time frame for the formation of European brown OREJUELA. 1998. Status of wild Andean bears and bear lineages. Molecular Ecology 16:401–413. policies for their management. Ursus 10:87–100. SA´ NCHEZ-MERCADO, A., J. FERRER-PARIS,E.YERENA,S. ———. 1999. Spectacled action plan. GARCI´A-RANGEL, AND K.M. RODRI´GUEZ-CLARK. 2008. Pages 157–168 in C. Servheen, S. Herrero, and B. Factors affecting poaching risk to vulnerable Peyton, editors. Bears: Status survey and conservation Andean bears Tremarctos ornatus in the Cordillera de action plan. IUCN/SSC Bear and Specialist Me´rida, Venezuela: Space, parks and people. Oryx Groups, Gland, Switzerland and Cambridge, UK. 42:437–447. PROCTOR, M., B. MCLELLAN,J.BOULANGER,C.APPS,G. SCHWARTZ, C.C., S.D. MILLER, AND M.A. HAROLDSON. STENHOUSE,D.PAETKAU, AND G. MOWAT. 2010. 2003. Grizzly bear Ursus arctos. Pages 556–586 in G.A. Ecological investigations of grizzly bears in Canada Feldhamer, B.C. Thompson, and J.A. Chapman, using DNA from hair, 1995–2005: A review of methods editors. Wild mammals of North America: Biology, and progress. Ursus 21:169–188. management, and conservation. John Hopkins Univer- RAYAN, D.M., AND S.W. MOHAMAD. 2009. The importance sity Press, Baltimore, Maryland, USA. of selectively logged forests for tiger Panthera tigris SETTLAGE, K.E., F.T. VAN MANEN, J.D. CLARK, AND T.L. conservation: A population density estimate in Penin- KING. 2008. Challenges of DNA-based mark–recapture sular Malaysia. Oryx 43:48–51. studies of American black bears. Journal of Wildlife RI´OS-UZEDA, B., H. GO´ MEZ, AND R.B. WALLACE. 2005. Management 72:1035–1042. Habitat preferences of the Andean bear (Tremarctos SHARMA, R.K., Y. JHALA,Q.QURESHI,J.VATTAKAVEN,R. ornatus) in the Bolivian Andes. Journal of Zoology GOPAL, AND K. NAYAK. 2010. Evaluating capture– 268:271–278. recapture population and density estimation of tigers in ———, ———, AND ———. 2007. A preliminary density a population with known parameters. Animal Conser- estimate for Andean bear using camera-trapping vation 13:94–103. methods. Ursus 18:124–128. SHERLEY, R.B., T. BURGHARDT, P.J. BARHAM,N.CAMP- ROBINSON, S.J., L.P. WAITS, AND I.D. MARTIN. 2009. BELL, AND I.C. CUTHILL. 2010. Spotting the difference: Estimating abundance of American black bears using Towards fully-automated population monitoring of DNA-based capture–mark–recapture models. Ursus African penguins Spheniscus demersus. Endangered 20:1–11. Species Research 11:101–111.

Ursus 22(1):47–64 (2011) 64 ANDEAN BEAR POPULATION ESTIMATES N Garshelis

SILVER, S.C., L.E.T. OSTRO, L.K. MARSH,L.MAFFEI, A.J. WALLACE, R.B., H. GOMEZ,G.AYALA, AND F. ESPINOZA. NOSS, M.J. KELLY, R.B. WALLACE,H.GO´ MEZ, AND G. 2003. Camera trapping for jaguar (Panthera onca)in AYALA. 2004. The use of camera traps for estimating the Tuichi Valley, Bolivia. Mastozoologı´a Neotropica jaguar Panthera onca abundance and density using 10:133–139. capture/recapture analysis. Oryx 38:148–154. WHITE, G.C., D.R. ANDERSON, K.P. BURNHAM, AND D.L. SOISALO, M.K., AND S.M.C. CAVALCANTI. 2006. Estimating OTIS. 1982. Capture–recapture and removal methods the density of a jaguar population in the Brazilian for sampling closed populations. Report LA-8787- Pantanal using camera-traps and capture–recapture NERP. Los Alamos National Laboratory, Los Alamos, sampling in combination with GPS radio-telemetry. New Mexico, USA. Biological Conservation 129:487–496. WOODS, J.G., D. PAETKAU,D.LEWIS, B.N. MCLELLAN,M. SPEED, C.W., M.G. MEEKAN, AND C.J.A. BRADSHAW. 2007. PROCTOR, AND C. STROBECK. 1999. Genetic tagging of Spot the match — Wildlife photo-identification using free-ranging black and brown bears. Wildlife Society information theory. Frontiers in Zoology 4, doi: Bulletin 27:616–627. 10.1186/1742-9994-4-2. WRIGHT, J.A., R.J. BARKER, M.R. SCHOFIELD, A.C. STETZ, J.B., K.C. KENDALL, AND C. SERVHEEN. 2010. FRANTZ, A.E. BYROM, AND D.M. GLEESON. 2009. Evaluation of bear rub surveys to monitor grizzly bear Incorporating genotype uncertainty into mark–recap- population trends. Journal of Wildlife Management ture-type models for estimating abundance using DNA 74:860–870. samples. Biometrics 65:833–840. TREDICK, C.A., AND M.R. VAUGHAN. 2009. DNA-based YERENA, E., J. PADRO´ N,R.VERA,Z.MARTI´NEZ, AND D. population demographics of black bears in coastal BIGIO. 2003. Building consensus on biological corridors North Carolina and Virginia. Journal of Wildlife in the Venezuelan Andes. Mountain Research and Management 73:1031–1039. Development 23:215–218. TROLLE, M., AND M. KE´ RY. 2003. Estimation of ocelot YOGANAND, K., C.G. RICE, A.J.T. JOHNSINGH, AND J. density in the Pantanal using capture–recapture anal- SEIDENSTICKER. 2006. Is the in India secure? ysis of camera-trapping data. Journal of Mammalogy A preliminary report on distribution, threats and 84:607–614. conservation requirements. Journal of the Bombay ———, A.J. NOSS, J.L.P. CORDEIRO, AND L.F.B. OLIVEIRA. Natural History Society 103:172–181. 2008. Brazilian tapir density in the Pantanal: A comparison of systematic camera-trapping and line- ZHANG, B., M. LI,Z.ZHANG,B.GOOSSENS,L.ZHU,S. transect surveys. Biotropica 40:211–217. ZHANG,J.HU, M.W. BRUFORD, AND F. WEI. 2007. Genetic viability and population history of the giant VITERI, M.P. 2007. Conservation genetics of Andean bears (Tremarctos ornatus) in northeastern Ecuador: Molec- panda, putting an end to the ‘‘evolutionary dead end’’? ular tools, genetic diversity and population size. Thesis, Molecular Biology and 24:1801–1810. University of Idaho, Moscow, Idaho, USA. ZUG, B. 2009. Individual identification and habitat use of ———, AND L.P. WAITS. 2009. Identifying polymorphic Andean bears on private lands in the Ecuadorian microsatellite loci for Andean bear research. Ursus Andes. Thesis, University of Wisconsin, Madison, 20:102–108. Wisconsin, USA. WAITS, L.P., AND D. PAETKAU. 2005. Noninvasive genetic sampling tools for wildlife biologists: A review of applications and recommendations for accurate data Received: 14 November 2010 collection. Journal of Wildlife Management 69:1419– Accepted: 14 February 2011 1433. Associate Editor: R. Harris

Ursus 22(1):47–64 (2011)