Andean Bear Use of the Epiphytic Bromeliad Tillandsia Fendleri at Quebrada El Molino, Venezuela
Total Page:16
File Type:pdf, Size:1020Kb
Andean bear use of the epiphytic bromeliad Tillandsiafendleri at Quebradael Molino,Venezuela Isaac R. Goldstein1 stick together, with bite marks on the white basal meristematicpart (Goldstein 1990). WildlifeConservation Society, 2300 Southern Although there are several species of epiphytic Boulevard,Bronx, New York10460, USA bromeliads Tillandsia available in forests between 2,400-3,000 m at QuebradaEl Molino (T. Key words: bromeliad, epiphyte, resource use, specta- compacta, T. T. tetrantha,T. Andeanbears cled bear, Tillandsia fendleri, Tremarctos ornatus, complanata, spiculosa), Venezuela at QuebradaEl Molino have been found to only eat the T. also feed on the fruit of Ursus15(1):54-56 (2004) largestspecies, fendleri. They a Lauraceae tree (Beilschmiedia sulcata). In paramo areas, bears feed on the terrestrial bromeliad Puya aristeguietae (Goldstein 1990). T. fendleri plants can Epiphytic bromeliads of the genus Tillandsia have reach 2 m in height (including the single central erect been reportedas importantin the Andean (spectacled) inflorescence) and 1 m in diameter (Smith 1971), bear diet throughoutthe bear's distribution(Mondolfi making them very conspicuous and easily distinguished 1971, 1989; Peyton 1980; Jorgenson and Rodriguez from smaller Tillandsia species. 1986; Rodriguez et al. 1986; Suairez1989; Goldstein During work at QuebradaEl Molino (Goldstein 1990, 1990, Rodriguez 1991; Eulert 1995). Epiphytic bro- 2002), I found that Andeanbear T.fendleri feeding sites meliads are locally abundantin Andean forests (Gentry were mostly located in tall, emergent trees with large and Dodson 1987). Because they are high in soluble canopies and in large trees at the forest-paramoedge, carbohydrates,fat, and protein, bromeliads are a nutri- where the abundanceof T. fendleri plants was higher tious food for bears (Goldstein 1990). (bromeliads/treex = 16.95 [SD = 10.89]; diameter at Bears often leave sign of their activitiesin areaswhere breastheight [dbh]x = 51.84 cm [SD = 22.80 cm]; tree they live; trails, scats, broken branches,claw marks on height x = 10.03 m [SD = 3.10]; n = 51). Because trees, and beds are common in bear habitats(Burst and climbing is energeticallycostly and the distributionand Pelton 1983). Andean bears are no exception, leaving abundance of bromeliads is clumped, I predicted that abundantsign especially while feeding on epiphytic and Andean bears would maximize expected rewardfor the terrestrialbromeliads (Peyton 1980, Suarez 1989). In effort of climbing trees with T. fendleri, preferentially Venezuela, sign of feeding on epiphytic bromeliadsby using trees with the greatestbromeliad loads. Andean bears has been found in mountain forests between 1,680 and 3,200 m and were the most abundant bear sign found in forests from 2,400 to 2,800 m Study area (Goldstein 1990). Quebrada El Molino, in the state of Trujillo, Spectacled bears usually produce 2 types of signs at Venezuela, covers approximately 105 km2 and varies epiphytic bromeliad feeding sites: claw marks on tree in elevation between 2,400 and 3,600 m. The area bark and piles of bromeliad leaves on the forest floor. includes 4 small basins radiatingfrom the intersectionof Claw marks are visible on tree bark from 40-60 cm Fila Llano Grandeand ParamoCastillejo. The topogra- fed above the ground to the branches where the bear phy is rugged, with steep slopes, rocky outcrops, and scars on the bromeliads. The length and depth of the precipices. depends on the bark of the tree species. Bears feed on The 3 characteristicvegetation types are high moun- the basal meristematicpart of the bromeliads, causing tain cloud forest, transition cloud forest-paramo, and the leaves to drop to the forest floor. Bears have been paramo.High mountaincloud forests are found on the found to eat 1-10 or more plants at each tree, and valleys and mountainhillsides between 2,400 and 3,200 Tillandsia fendleri plants have around 50 leaves. m and are composed of small to mediumtrees (5-10 m in Depending on the number of bromeliads eaten, leaves height)with few emergingtrees > 15 m high. Withinhigh can form a conspicuous carpet on the forest floor. mountain cloud forests are patches of undisturbedold Moreover, usually the central bunch of younger leaves growth forest, disturbed or regenerating forest, and patchesdominated by Alnus spp. Undisturbedold growth is characterizedby multi-stratifiedforest with under- the '[email protected] growth that is either open or closed (depending on 54 SHORTCOMMUNICATIONS 55 presence and relative abundanceof the bamboo Rhipi- was 3.16/tree (SD = 7.10). Andean bear feeding sign on docladum germinatum),many small and medium sized epiphytic bromeliads was observed at 13 trees. Sixty trees, and few big emergent trees. The most common seven T.fendleri plants were found eaten, and 263 were emergenttree species are Podocarpus oleifolius, Beilsh- still present in the canopy of the 123 trees observed miedia sulcata, and Clusia spp., marked by irregular, (20.3% of the availableplants were consumed).Number horizontalbranches and high loads of epiphyte vegeta- of T.fendleri per tree was positively correlatedwith tree tion. Disturbed or regeneratingforest patches indicate dbh (r, = 0.455, P < 0.001) and tree height (r, = 0.625, past tree cutting or regeneratingpastures or croplands. P < 0.001). For trees lacking evidence of use by bears, These are characterizedby a homogenous canopy of the number of T. fendleri per tree was similarly Miconia spp., Senecio spp., and Weinmanniaspp. trees, correlatedwith dbh (r, = 0.532, P < 0.01) and height with no emergenttrees. Alnus-dominatedforest patches (r = 0.401; P < 0.01). However, for trees used by bears, are homogeneous and result from natural landslides, correlationsbetween the numberof T. fendleri per tree usually near river beds. The transition cloud forest- and dbh (r, = 0.401, P = 0.175) and height (r = 0.208, paramovegetation type is found between 3,100-3,300 m P = 0.495) were not significant.Trees showing use by and is composed of trees <6 m tall with a clear bears had significantlygreater dbh (Mann-WhitneyU= dominance of Clusia spp. trees. The pairamovegetation 1384.5, P < 0.05), height (U = 1432.5, P < 0.05), and type is basically treeless and dominatedby graminoids number of bromeliads (U= 1485.0, P < 0.05) than and species of the Espeletia group. unused trees. Methods Discussion I established 10 transects(200 x 5 m) in the ecotone The results of this small-scale study support the between forest and paramo vegetation types during 3 hypothesisthat Andean bears use trees with higherloads visits to the study area, July 1997-January 1998. The of epiphytic bromeliadsT. fendleri, which happento be transects' initial points were subjectively selected the larger trees. T. fendleri appears to be a canopy throughout the 7 km of the paramo-forest ecotone specialist species, being more abundant in higher within the QuebradaEl Molino watershed,reflecting the emergent trees. Gentry and Dodson (1987) found that feasibility of walking downslope regardlessof compass Bromeliaceaespecies are usually habitat-restricted,with bearing. Height, dbh, and number of T. fendleri were different species restricted to the understory, middle measured for all trees >10 cm dbh. Two observers story, or canopy of the trees. Specialization to open tallied T. fendleri in the canopy of the trees, and the and wind exposed habitatsis common among members results were averaged. In trees showing sign of bear of the subfamily Tillandsoidea having plumed, wind in the we feeding canopy, counted the bromeliadleaves propelled seeds. Because wind is the main dispersal on the forest floor and divided 50 by (averagenumber of agent, the probabilityof seeds landing and establishing leaves bromeliad per plant), approximatingthe number themselves is higher where seeds are exposed to wind of the bromeliad consumed plants in that tree. than in sites protectedfrom wind. Tall emergent trees, Because and the dbh, height, number of bromeliads trees at the forest-paramoedge, and solitarytrees are all tree were not distributed per normally (Kolmogorov- wind-exposed. Moreover, emergent and solitary trees Smimov P < all test, 0.05), statistical tests were non- collect more water from horizontal precipitationthan To between parametric. quantifyrelationships numberof trees inside the forest (Vogelmann 1973), and epiphytes T. and tree I used fendleri size, Spearman'srank order are particularlyable to take advantage of horizontal correlation.I correlation performedseparate analyses on precipitation,being more abundantin the highest trees all trees, trees used and treesnot used by bears, by bears. (Nadkarmi1984). Thus, if humidityand wind exposure I used the test to 2-sample Mann-Whitney comparesizes are importantto the life cycle of T. fendleri, a positive of used and not used trees. correlation between the height of the trees and the numberof T. fendleri is expected. I documented Results Although only 13 trees used by bears, the numberof T. fendleri used by bears implied by this A total of 123 trees >10 cm dbh were sampled.Mean sample is quite high. I sampled only 1 ha within the dbh was 28.19 cm = (SD 22.61 cm), mean height was 105 km2 study area;within the trees used for bromeliad 6.94 m = 2.28 (SD m), and mean numberof T.fendleri feeding there was a mean of 5.15 eaten bromeliads.At Ursus 15(1):54-56 (2004) 56 SHORTCOMMUNICATIONS least half of the study area is forested (52 km2); JORGENSON, J., AND J.V. RODRIGUEZ.1986. A preliminary extrapolatingthe number of trees used