Radiotelemetry Data Research Articles

Home range sizes of wildcats (Felis silvestris) and feral domestic cats (Felis silvestris f. catus) in a hilly region of Hungary

Summary The most important factor concerning wild cat populations is the loss of habitat. Therefore, it is necessary to assess the size of the home ranges of wild and domestic cats along with the features of these areas (vegetation, elevation, proximity to human settlement, etc.). A total of 16 wildcats and 19 domestic cats were caught and fitted with radio collars within the period between 1989–1993. It was possible to analyze the radiotelemetry data of 4 wildcats and 3 domestic cats. It resulted that the wildcats occupied larger home ranges than the domestic cats, however, there were exceptions. Home range size variability was extensive in both species. The males occupied larger areas than the females. This was most likely due to the reproductional wandering of males into female home ranges. Also the overlap between the home ranges of males was larger than that of females. However, there were very small overlaps between the core areas. No cats used the same sites at the same time. This indicates that the home ranges of cats exist only in space and time as well. Although these animals are solitary, there was some indication that hierarchy exists between males.

Consumption of salmon by Alaskan brown bears: a trade-off between nutritional requirements and the risk of infanticide?

The risk of infanticide may alter foraging decisions made by females, which otherwise would have been based on nutritional requirements and forage quality and availability. In systems where meat resources are spatially aggregated in late summer and fall, female brown bears (Ursus arctos) would be faced with a trade-off situation. The need of reproductive females to accumulate adequate fat stores would likely result in a decision to frequent salmon streams and consume the protein- and lipid-rich spawning salmon. In contrast, aggregations of bears along salmon streams would create conditions of high risk of infanticide. We investigated consumption of salmon by brown bears on Admiralty and Chichagof Islands in Southeast Alaska from 1982 to 2000 using stable isotope analysis and radiotelemetry. While nearly all males (22 of 23) consumed relatively large amounts of salmon (i.e., >10% relative contribution to seasonal diet), not all females (n=56) did so. Five of 26 females for which we had reproductive data, occupied home ranges that had no access to salmon and thus did not consume salmon when they were mated or accompanied by young. Of females that had access to salmon streams (n=21), all mated individuals (n=16) had delta(15)N values indicative of salmon consumption. In contrast, 4 out of 16 females with cubs avoided consuming salmon altogether, and of the other 12, 3 consumed less salmon than they did when they were mated. For 11 of 21 females with access to salmon streams we had data encompassing both reproductive states. Five of those altered foraging strategies and exhibited significantly lower values of delta(15)N when accompanied by young than when mated, while 6 did not. Radiotelemetry data indicated that females with spring cubs were found, on average, further away from streams during the spawning season compared with females with no young, but both did not differ from males and females with yearlings and 2-year-olds. Females with young that avoided salmon streams were significantly lighter indicating that female choice to avoid consumption of salmon carries a cost that may translate to lower female or cub survivorship. The role of the social hierarchy of males and females, mating history, and paternity in affecting the risk of infanticide and foraging decisions of female brown bears merit further investigation.

IDENTIFYING BREEDING HABITAT FOR THE IBERIAN LYNX: INFERENCES FROM A FINE‐SCALE SPATIAL ANALYSIS

Empirical landscape habitat models are valuable tools for addressing species conservation issues in heterogeneous landscapes. These have been particularly useful for animal populations requiring extensive areas, like large mammalian carnivores. Although models are scale‐dependent, they are often based exclusively on coarse‐grained information on landscape structure and species distribution. However, accurate discrimination and quality assessment of breeding habitats may require more detailed information on both individuals' habitat use and landscape patterns.Here, we modeled breeding habitat for the Iberian lynx (Lynx pardinus), an endangered specialist of Mediterranean native ecosystems, in a human‐dominated landscape. For this, we used radiotelemetry data on breeding individuals and fine‐grained landscape variables obtained over an areal extent encompassing an entire lynx metapopulation. In addition, we restricted the study domain to areas of potential presence within this extent based on previous habitat analyses. This fine‐scale landscape analysis and design aimed to test whether it was possible to obtain detailed inferences on breeding habitat distribution and quality within more coarsely defined habitats that only indicate where the species can be found. Landscape composition factors (i.e., fine‐grain variables of landscape pattern within territories) and context factors (variables on geographical location of territories) were combined in different candidate habitat models. These reflected various hypotheses involving territory presence and size, including vegetation structure characteristics, landscape heterogeneity and complexity, favorable prey habitat, and human disturbance. The best approximating model for territory presence included only one landscape composition variable, namely, density of ecotones between scrubland and pastureland, which favored presence. For this model, classification accuracy was >80% in 94% of cross‐validation tests performed. The best approximating model for territory size explained 65% of the variation in size and included also density of ecotones, plus mean coverage of tall shrubs within territories; both correlated negatively with size. We show how these predictors are also related to prey density and suggest that this low‐hierarchical‐level relationship between the prey and patterns of the landscape can be interpreted as a likely mechanism explaining breeding habitat distribution and quality at a higher level. Our approach shows the potential of fine‐scale landscape analysis to understand patterns of breeder abundance, and to better manage population viability of threatened species in human‐populated areas.

Age-specific survival of Florida's Crested Caracaras

Abstract To evaluate the status of Florida's threatened Crested Caracara (Caracara cheriway) population, knowledge of life history parameters and population dynamics is required. Using the program MARK, I estimated survival rates for adult and juvenile Crested Caracaras from breeding areas in southcentral Florida during the period February 1994 through March 2000. Annual adult survival probabilities estimated from resightings of banded individuals and derived using model averaging were 0.876 for males and 0.906 for females. Results of modeling monthly juvenile caracara survival from radiotelemetry data identified differences in survival for two age groups: juveniles (up to age 3 years) and adults (age >3 years). Using this age-effects model, monthly survival for juvenile caracaras was estimated to be 0.970, indicating an annual survival rate of 0.694 and an overall probability of 0.334 of an individual surviving to age 3, when it could potentially become a breeder. Juvenile caracaras experienced periods o...

Effects of a Stationary GPS Fix-Rate Bias on Habitat-Selection Analyses

Global Positioning System (GPS) radiotelemetry, as a data collection technique, currently suffers from largely unknown fix-rate biases that result in lost data due to environmental factors, which could lead to wrongful research conclusions. Estimates of bias within uncorrected data from free-ranging animals have not been reported. I tested the effects of a known fix-rate bias model on habitat-selection conclusions by correcting habitat-use data for mule deer (Odocoileus hemionus) and random data, and comparing outcomes to those of the original uncorrected data. In all cases, corrected and uncorrected data yielded similar habitat-selection conclusions. I attributed this to the model, which accounted for <10% data loss in all cases. Due to this small amount of explained data loss, I suggest that animal behavior (e.g., bedding vs. moving) is a large source of unexplained data loss from free-ranging animals and must be accounted for to best overcome biases in GPS radiotelemetry data.

Population Structure, Growth, and Age Estimation of Spotted Turtles, &lt;em&gt;Clemmys guttata&lt;/em&gt;, Near their Northern Limit: an 18-Year Follow-up

Mark-recapture and radio telemetry data were collected from 1999-2001 on a population of Spotted Turtles, Clemmys guttata, near their northern limit in Ontario, Canada, and compared with a similar study conducted from 1983-1986. In 1999-2001, 32 Spotted Turtles were caught, 27 of which were adults. Females outnumbered males 3.5:1. The carapace length (CL) of adult males (108.5 ± 5.7 mm) and adult females (106.4 ± 6.0 mm) did not differ significantly. A minimum of 22% of juveniles and 40% of adults marked in 1983-1984 were still alive in 1999-2001. Adult population size was estimated to be 45 (95% confidence interval: 34-78). The average size of individual adults marked in 1983 (105.6 + 6.0 mm CL) and recaptured during 1999-2001 (106.3 + 5.9 mm CL) did not differ significantly. There was also no signficant difference in the number of growth lines on adults caught in 1983 (12.9 ± 2.0) and 1999 (14.0 ± 2.3). Based on the number of growth lines in 1983, these turtles averaged a minimum of 28.9 years in 1999, but actual ages are likely greater. Summer range length averaged 233 m (range: 200-275 m), while total range length averaged 327 m (range: 275-375 m).

The Ranging Behaviour and Habitat Use of Rabbits on Farmland and Their Potential Role in the Epidemiology of Paratuberculosis

Grazing herbivores avoid grass swards contaminated with faeces as the ingestion of faeces is a common route of micro- and macro-parasite transmission. The recent novel finding that herbivores do not avoid grass swards contaminated with rabbit faeces suggests that disease risk posed to herbivores by rabbits is determined by rabbit ranging and excretory behaviour. Using as a case study rabbits and the risk Mycobacterium avium sub-species paratuberculosis ( M. a. paratuberculosis) poses to cattle, the interaction between rabbits and grazing pasture was studied on an infected farm in the east of Scotland in spring and autumn 2000. Radio telemetry, burrow surveys and faecal pellet count data were collected on two areas (Areas 1 and 2) of the farm with different habitat mosaics, to study the potential effects of season and habitat on the spatial distribution of rabbits faeces and thus disease in the environment. Twenty one rabbits were radio tracked and a total of 902 fixes collected. Mean home range sizes (100% minimum convex polygons) were between 2.0 and 7.1 ha per rabbit per season. Home ranges were significantly larger in spring, and in Area 1 which had more moor and woodland and less rough pasture. Rabbits used rough pasture most in Area 1 and gorse scrub in Area 2. In both areas, significantly more burrows were located in gorse scrub than in any other habitat. Most faecal pellets were deposited on the moorland habitat of Area 2 in autumn. In habitats to which grazing livestock had access, the mean rate of faecal deposition was 8571 pellets per ha per day. The greatest risk of disease transmission occurred in habitats of poor grazing quality (e.g., gorse scrub) which were used by rabbits for burrowing and thus contained high concentrations of faeces. The findings of the study are discussed in relation to management practices aimed at reducing disease risk to livestock, including the fencing of scrub and the reduction of rabbit population size to prevent expansion of rabbit burrows from scrub into grazing pastures.

Using the bootstrap and fast Fourier transform to estimate confidence intervals of 2D kernel densities

Kernel density estimators are often used to estimate the utilization distributions (UDs) of animals. Kernel UD estimates have a strong theoretical basis and perform well, but are usually reported without estimates of error or uncertainty. It is intuitively and theoretically appealing to estimate the sampling error in kernel UD estimates using bootstrapping. However, standard equations for kernel density estimates are complicated and computationally expensive. Bootstrapping requires computing hundreds or thousands of probability densities and is impractical when the number of observations, or the area of interest is large. We used the fast Fourier transform (FFT) and discrete convolution theorem to create a bootstrapping algorithm fast enough to run on commonly available desktop or laptop computers. Application of the FFT method to a large (n>20,000) set of radio telemetry data would provide a 99.6% reduction in computation time (i.e., 1.6 as opposed to 444 hours) for 1000 bootstrap UD estimates. Bootstrap error contours were computed using data from a radio-collared polar bear (Ursus maritimus) in the Beaufort Sea north of Alaska.

LANDSCAPE CONSERVATION AND REGIONAL PLANNING FOR THE FLORIDA PANTHER

The need for regional planning is increasingly important for effective Florida panther (Puma concolor coryi) (Bangs 1898) conservation and is essential for protecting enough habitat in South Florida to ensure a viable population. We used two decades of radio telemetry data and geographic infor- mation system (GIS) software to develop a regional blueprint for landscape restoration that enhances dispersal, facilitates population colonization, and could be the basis for future land use decisions in the range of the endangered Florida panther. We identified 923,576 ha of forests in an 18-county study area that is a barrier-rich patchwork of land uses. A least cost path analysis simulated natural colonization events and can be used to identify landscape linkages and conservation networks for the panther. Our analysis of planned development permits suggests that large-scale land protection must happen quickly. The alternatives are managing an isolated, heavily managed population or large- scale landscape restoration that is probably economically unfeasible.

A Comparison of Home Ranges of Two Species of Peromyscus Using Trapping and Radiotelemetry Data

A Comparison of Home Ranges of Two Species of Peromyscus Using Trapping and Radiotelemetry Data

A COMPARISON OF HOME RANGES OF TWO SPECIES OF PEROMYSCUS USING TRAPPING AND RADIOTELEMETRY DATA

We investigated differences between sizes of home ranges using trapping and radiotelemetry data for syntopic Peromyscus boyliiand P. truei. Sizes of home ranges were calculated from the minimum convex polygon of trap locations and radiotelemetry locations and compared between individuals. The 2 estimates of home-range size were significantly correlated, although on an average trapping home ranges were significantly smaller than sizes of radiotelemetry home ranges. Home-range sizes from radiotelemetry were inversely correlated with conspecific density, but home-range sizes from trapping were not. Thus, at low density, radiotelemetry home ranges were significantly larger than trapping home ranges, but at high density there was no difference between radiotelemetry and trapping home ranges. These results indicate that radiotelemetry results in larger estimates of home-range size, particularly at lower densities of conspecifics. The largest size estimates of home ranges were from a combination of radiotelemetry and trapping data.

Optimal allocation of sample sizes between regular banding and radio-tagging for estimating annual survival and emigration rates

Many authors have shown that a combined analysis of data from two or more types of recapture survey brings advantages, such as the ability to provide more information about parameters of interest. For example, a combined analysis of annual resighting and monthly radio-telemetry data allows separate estimates of true survival and emigration rates, whereas only apparent survival can be estimated from the resighting data alone. For studies involving more than one type of survey, biologists should consider how to allocate the total budget to the surveys related to the different types of marks so that they will gain optimal information from the surveys. For example, since radio tags and subsequent monitoring are very costly, while leg bands are cheap, the biologists should try to balance costs with information obtained in deciding how many animals should receive radios. Given a total budget and specific costs, it is possible to determine the allocation of sample sizes to different types of marks in order to minimize the variance of parameters of interest, such as annual survival and emigration rates. In this paper, we propose a cost function for a study where all birds receive leg bands and a subset receives radio tags and all new releases occur at the start of the study. Using this cost function, we obtain the allocation of sample sizes to the two survey types that minimizes the standard error of survival rate estimates or, alternatively, the standard error of emigration rates. Given the proposed costs, we show that for high resighting probability, e.g. 0.6, tagging roughly 10-40% of birds with radios will give survival estimates with standard errors within the minimum range. Lower resighting rates will require a higher percentage of radioed birds. In addition, the proposed costs require tagging the maximum possible percentage of radioed birds to minimize the standard error of emigration estimates.

Influences of winter supplemental feeding on the energy balance of white-tailed deer fawns in New Hampshire, U.S.A.

The purpose of this study was to determine the influence of supplemental feeding on the energy balance of white-tailed deer (Odocoileus virginianus) in northern New Hampshire, U.S.A., during January–March 1997. We measured the field metabolic rate (FMR) and energy balance of 10 (5 males and 5 females) supplementally fed wild fawns with doubly labeled water over 19–21 days. We tested whether percent acid detergent fiber (ADF) and percent neutral detergent fiber in deer fecal samples predicted the proportion of supplemental feed (pelleted concentrate) in the diet of deer. The mean FMR of fawns was 758.4 kJ·kg–0.75·d–1 (range = 535.9–1032.8 kJ·kg–0.75·d–1), or 2 × their basal metabolic rate (BMR). The mean FMR of male fawns was &gt;30% higher than that of female fawns. Percent body fat (12.1 ± 1.4% (mean ± SE)) and mass loss (3.0 ± 0.9%) varied among fawns, suggesting that an individual high FMR was not detrimental to energy balance and was related to availability of feed. We estimated that a high FMR (&gt;2 × BMR) could be maintained only if fawns consumed about 1 kg of supplemental feed daily. Radiotelemetry data indicated that the number and juxtaposition of feeding sites in an area probably influenced home range and activity of deer. Percent ADF in feces provided the best prediction of percent grain in the diet (% grain = –0.048(% ADF)2 + 1.523(% ADF) + 96.467; r2 = 0.69) and was useful for identifying populations consuming supplemental pelleted concentrate. Biologists should expect that the influence of winter feeding on energy balance and survival will vary according to the interrelationships of deer density, food availability, and winter severity at feeding sites.

Fish individual-based numerical simulator (FINS): a particle-based model of juvenile salmonid movement and dissolved gas exposure history in the Columbia River basin

This paper describes a numerical model of juvenile salmonid movements in the Columbia and Snake Rivers. The model, called the Fish Individual-based Numerical Simulator or FINS, employs a discrete, particle-based approach to simulate the movements and history of exposure to dissolved gases of individual fish. FINS is linked to a two-dimensional (vertically-averaged) hydrodynamic simulator that quantifies local water velocity, temperature, and dissolved gas levels as a function of river flow rates and dam operations. Simulated gas exposure histories can be input to biological mortality models to predict the effects of various river configurations on fish injury and mortality due to dissolved gas supersaturation. Therefore, FINS serves as a critical linkage between hydrodynamic models of the river system and models of biological effects. FINS model parameters were based on observations of individual fish movements collected using radiotelemetry methods during 1997 and 1998. A quasi-inverse approach was used to decouple fish swimming movements from advection with the local water velocity, allowing inference of time series of non-advective displacements of individual fish from the radiotelemetry data. Statistical analyzes of these displacements confirm that strong temporal correlation of fish swimming behavior persists in some cases over several hours. A correlated random-walk model was employed to simulate the observed behavior, and parameters of the model were estimated that lead to close correspondence between predictions and observations.

Estimating survival from joint analysis of resighting and radiotelemetry capture-recapture data for wild animals

Biologists often use more than one marking technique in wildlife studies. For each of the mark types, it is common to conduct a separate analysis of the recapture data to estimate parameters of interest, such as survival rates. Two data types that can be used in estimating survival rates are resighting and radiotelemetry data. The Cormack-Jolly-Seber model is commonly used to analyze the resighting data, while the Kaplan-Meier product limit estimator, modified for staggered entry of animals, is used to analyze the radi otelemetry data. In a study where some animals receive two types of tags and others receive just one tag type, the separate Cormack-Jolly-Seber and Kaplan-Meier analyses do not exploit all of the information in the combined data sets. In this article, we propose a model and likelihood for the combined analysis of resighting and radi otelemetry data. In comparison with the separate analyses, this richer model provides more information about the biology and sampling processes. For example, the richer model permits assessment of assumptions required by the separate analyses and allows estimation of additional parameters. We apply the model to annual resighting and monthly telemetry data from a population of snail kites in Florida. The snail kite is a threatened species of bird in the United States, and our results on survival are very important. In this example, all birds are marked using leg bands and some of them receive radios.

Modeling Snake Microhabitat from Radiotelemetry Studies Using Polytomous Logistic Regression

-Multivariate analysis of snake microhabitat has historically used techniques that were derived under assumptions of normality and common covariance structure (e.g, discriminant function analysis, MANOVA). In this study, polytomous logistic regression (PLR), which does not require such assumptions, was developed to model habitat use patterns from snake radiotelemetry data. Two case study examples were used: the copperhead (Agkistrodon contortrix) in a forested habitat in Chesapeake, Virginia, and the cottonmouth (Agkistrodon piscivorus piscivorus) in a wetland habitat in Virginia Beach, Virginia. The model was developed based on grid-cell counts of the habitat used by adult snakes that were radiotracked for a minimum of one activity season (early spring through late fall). In addition to PLR modeling, a nonparametric MANOVA procedure based on ranked data from presence versus random sites was developed for comparative purposes. Although results were similar in terms of variables chosen by the models, nonparametric procedures lack predictive power, and conclusions drawn from them are sometimes questionable or difficult to interpret. Polytomous logistic regression provides a useful alternative to traditional modeling approaches that require ancillary data from random sites; PLR requires data only from sites used by snakes, and was developed based on use-intensity categorization. Suggestions for model implementation (e.g, in a

Applying a variable-shape spatial filter to map relative abundance of manatees (Trichechus manatus latirostris)

Presented is a modified spatial filter, called a variable-shape filter, that was used to transform a map of point locations of Florida manatees (Trichechus manatus latirostris) to a contoured surface illustrating relative abundance. Rather than having a fixed polygon shape and size as conventional filters do, this method preserves constant filter area but conforms polygon shape to include only the desired land-cover categories–in this study, water. Variable-shape filter polygons are formed by starting with the cell that the point is located in and then adding the nearest contiguous cells of the desired land-cover category to the polygon until the area requirement is reached. Surfaces generated using the variable-shape filter were compared to those created with a conventional, circular, fixed-shape filter. Four filter sizes, based on an analysis of manatee hourly travel rates estimated from satellite and radio telemetry data, were used. Filter sizes, defined in terms of a circle's radius, were 125 m, which was the 25th percentile of the cumulative manatee travel-rate distribution; 325 m, the 50th percentile; 800 m, the 75th percentile; and 3,950 m, the 99th percentile. The fixed-shape and variable-shape filters differed principally in how their results were influenced by land. The variable-shape filter, programmed to maintain constant area, estimated animals to occur farther from shore than the fixed-shape filter did. Fixed-shape filter polygons were occasionally divided by land barriers, such as peninsulas, resulting in calculations of relative abundance estimates that were near the visual sighting in terms of euclidian distance but far in terms of manatee travel. The variable-shape filter was preferable primarily because it was more sensitive to manatee ecology: only cells contiguous to the animal's mapped location were included in the filter calculations.

Activity and Movement Patterns of Polar Bears Inhabiting Consolidated versus Active Pack Ice

We investigated the influence of ice conditions on activity and movement patterns of polar bears in the Canadian-West Greenland Arctic. We used radiotelemetry data gathered over 11 years (1989-99) from 160 adult female polar bears to test for differences in movement and activity of bears inhabiting active ice and consolidated ice. Bears inhabiting active ice moved more than those inhabiting consolidated ice (12 versus 8 km/day), but their activity throughout the year did not differ (bears of both groups were active for 21% of the day). Differences in activity and movement of bears in the two study areas appeared to be related to differences in predominant ice conditions and presumed prey availability. Seals, particularly juveniles, are most plentiful in spring and summer, when polar bears moved more and were most active. During winter, when juvenile seals were less available in consolidated ice areas, bears in that habitat were less active and moved less than bears in active ice areas. Polar bears have evolved flexible patterns of seasonal activity, movements, and facultative den use as adaptations to different sea-ice environments.

A note on silent ovulation identified by using radiotelemetry for estrous detection

This study was undertaken to determine the occurrence of silent ovulation (SO) in high-producing, postpartum (PP) Holstein cows. Nineteen high-producing, multiparous cows received transmitters (Heatwatch®, DDX, Denver, CO) for radiotelemetry (RT) and began visual estrus detection (VED) routine for 1/2 h twice daily, on day 10 PP. Blood samples were obtained every other day commencing on day 10 PP, for RIA of progesterone (P 4). Following P 4 assay, RT and VED data were evaluated at time periods coinciding with ovulation to determine whether standing estrus had occurred. First ovulation occurred at 29.1±3.0 days PP (range 17–69 days) compared to first mount at 39.6±4.0 days PP (range 17–69 days) and at 47.6±3.6 days PP (range 24–77 days) for RT and VED, respectively. VED indicated 94.7% of first ovulations as SO while the inclusion of RT with VED indicated that only 42.1% of first PP ovulations were true SO. During second PP ovulations ( n=16), VED indicated 50% SO while the inclusion of RT indicated only 12.5% true SO. During third and fourth PP ovulations ( n=15), VED indicated 33.3% SO while the inclusion of RT indicated only 6.7% true SO. When RT was associated with ovulation, there were fewer mounts at first PP ovulations compared to subsequent ovulations (1.1±0.1 vs. 3.8±0.7, P<0.0005) and lengths of individual mounts were shorter (2.6±0.5 vs. 3.4±0.3 s, P<0.05). Previous studies have indicated the majority of first PP ovulations to be SOs. However, it appears more likely that while a substantial number are SOs, more are associated with a reduced expression of behavioral estrus.

Home Range, Population Density, and Food Resources of Agouti paca (Rodentia: Agoutidae) in Costa Rica: A Study Using Alternative Methods1

ABSTRACTField studies of elusive mammals such as Agouti paca are difficult, and results are often biased and incomplete because the animals are nocturnal and avoid traps. By studying an A. paca population in a Pacific lowland forest of Costa Rica, we developed indirect observational methods to estimate population density and food resource utilization. To estimate population density, we divided the number of A. paca burrows found in a 12‐ha census area by the average number of burrows utilized per individual from radiotelemetry data. We compared this estimate to independent population estimates obtained using two transect methods. Agouti paca density was estimated at 93 individuals/km2 based on burrow number, and between 67 to 70 individuals/km2 using transect methods. This discrepancy corroborates previous studies suggesting that transect methods underestimate population density. To identify food resources consumed by the species, we collected fruits and seeds with unambiguous A. paca teeth marks. The only animal in Central America with an incisor width similar to A. paca. is Dasyprocta variegata; in a comparison of both species, however, we found that A. paca had a significantly larger incisor width and that teeth marks of 4 mm or wider unambiguously indicated handling by this species. By applying this technique, food resources utilized by A. paca can be identified more easily than by direct observation. At our study site, we identified 33 plant species consumed by A. paca; 61 percent are new records for this animal. Our results suggest that A. paca plays a major, although not fully understood, ecological role in seed predation and dispersal.