Abstract

Animals should use habitats that increase their chances of survival and reproductive output; thus, information on habitat use can be useful both from an ecological and a conservation perspective. Using global positioning system tracking collars, we studied habitat use by a pair of Chrysocyon brachyurus (maned wolves) living in a cerrado-Atlantic forest transition zone in Brazil. During our study, the female had a litter of pups, and we were able to evaluate how reproduction affected habitat use. To investigate habitat use, we mapped our telemetry fixes onto vegetational and geographical data obtained from satellite images of our study site. We investigated habitat use by maned wolves in relation to vegetation and land cover, proximity to buildings, proximity to roads, elevation, and distance to water sources. Maned wolves showed a strong preference for rocky fields, which are a subhabitat of cerrado, and a strong avoidance of Atlantic rain forest. Maned wolves also were associated significantly and positively with buildings and roads at night, likely because of supplemental feeding and ease of locomotion. Elevation was confounded with habitat type and the location of supplemental feeding; therefore, no firm conclusions could be drawn about this factor. The female remained ≤250 m from water when she was lactating. Although maned wolves can be behaviorally flexible in terms of habitat use, they show a strong preference for open habitats such as cerrado that is especially pronounced during a reproductive phase.

The maned wolf (Chrysocyon brachyurus) is considered to be a species typical of savanna-type habitats, such as the Brazilian cerrado, and is considered to be a flagship species for this biome. This species displays many anatomical adaptations for survival in this type of habitat. The maned wolf has long legs relative to its body length, permitting it not only to travel quickly but also giving it a high vantage point from which to hunt rodents and ground-dwelling birds (Dietz 1984, 1985). However, the species appears to be able to acclimate to a variety of habitat types. For example, individuals have been observed in the Atlantic rain-forest zone in Brazil, where they are probably taking advantage of cattle pastures created by deforestation (Santos et al. 2003). The Atlantic forest currently is composed of a mosaic of small forest fragments distributed in agricultural land (Tabarelli et al. 2005), with only 7% of its original distribution remaining.

At present, there have been no studies of habitat use by maned wolves. This lack of data probably reflects the difficulty in observing the species in the wild, and that the majority of study sites, both historical and actual, are located in cerrado habitat. Studies conducted at sites dominated by a single type of habitat, such as the classic study of the ecology of maned wolves by Dietz (1984) in Serra da Canastra National Park, Brazil, do not allow true assessment of habitat use because comparisons showing avoidance or selection of a variety of habitat types are not possible. Broomhall et al. (2003), studying another open-habitat species, the cheetah (Acinonyx jubatus), stressed the importance of studying such species in a variety of habitat types to determine a habitat preference. Studies of alien invader species have shown that where we encounter a species is not necessarily where it would thrive in terms of reproduction and survival (Hulme 2006).

Conservation biologists are increasingly turning to analysis and interpretation of satellite imagery to help them in their studies to conserve animal species. The interpretation and analysis of satellite images allows researchers to identify and classify vegetation cover and land use without needing to visit a particular location, thereby greatly speeding up environmental research. In one recent study, Teixeira et al. (2006) were able to identify 21 types of vegetation cover and land use in a wide variety of widely distributed Brazilian environments from satellite images. This use of satellite images can become particularly powerful when it is combined with studies of animals that are fitted with global positioning system (GPS) tracking collars to monitor habitat use because GPS collars provide data on animal locations with a relatively high precision (4-15 m depending on habitat density—Coelho et al. 2007; Wing et al. 2005). Radiotelemetry studies using traditional radiocollars and triangulation can be relatively inaccurate for species that travel large distances each day, such as maned wolves (de Melo et al. 2007). For example, a radiotelemetry study on European wolves (Canis lupus) reported an average error of >200 m on each location determined by telemetry (Theuerkauf and Jedrzejewski 2002); such an error would result in highly inaccurate estimates of habitat use (Garshelis 2000).

To preserve a species, it is necessary to understand how it interacts with its environment, and which habitats within its environment are important to its survival (Garshelis 2000). Increasingly, animals are living in environments that have been modified by humans, with urban wildlife being an extreme example of this (Williams et al. 2006). However, even in nature reserves animals are frequently subjected to a modified environment where roads and trails provide access to ecotourists (Trombulak and Frissell 2000). Frequently, nature reserves house facilities such as hotels and other infrastructure to support ecotourism, and may deliberately attract wildlife for tourists to observe by supplemental feeding (Silva and Talamoni 2004). At our study site, this type of supplemental feeding occurred on a daily basis for wild maned wolves (Silva and Talamoni 2004). A study on the effects of supplemental feeding on white-tailed deer (Odocoileus virginianus) showed that it did not affect home-range size but did reduce size of core areas by 50% (Cooper et al. 2006).

In this study, we used GPS tracking collars to examine habitat use by wild maned wolves both during and outside a reproductive event. We chose a study site that was in a transition zone between cerrado (considered typical habitat for maned wolves) and Atlantic rain forest. The use of habitat in this type of environment will allow evaluation of the importance of these different types of habitat to maned wolves.

Materials and Methods

All procedures reported in this paper are in compliance with all relevant and applicable state and federal laws in Brazil. The subject of our study was a pair of wild maned wolves. During the middle of this study the female gave birth to 1 pup (in the middle of June 2001 in the dry season), and made her den inside a hollowed-out termite mound. This study was conducted in the Reserva Particular do Patrimônio Natural Santuário da Serra do Caraça (Private Natural Heritage Reserve [20°05′S, 43°29′W]), located in Minas Gerais State, southeastern Brazil, in the southern portion of the Espinhaço Mountain range. The area of the reserve was 10,188 ha, with elevations ranging from 850 to 2,072 m above sea level. The Espinhaço Mountain range is a mountainous complex that delimitates a zone of contact between the cerrado and the Atlantic forest in its southern portion, and transition zones of cerrado, Atlantic forest, and caatinga (tropical deciduous forest) in its central and northern edges (Giulietti and Pirani 1988; Giulietti et al. 1997). In the center of the reserve is a hotel where supplemental feeding (approximately 2 kg of meat per day, which is approximately 30% of the daily intake of food of an adult maned wolf) of the maned wolves occurred on a nightly basis between 1900 and 2200 h. The reserve and the supplemental feeding have been described in detail by Silva and Talamoni (2004).

Our study was conducted from August 2000 to February 2002. The male maned wolf was captured on 19 July 2000 and the female was captured on 7 August 2000 in a standard wooden trap designed for the species (see Dietz 1984). After capture, maned wolves were anesthetized and fitted with GPS tracking collars (model Simplex; Televilt, Lindesburg, Sweden) under the supervision of a qualified veterinary surgeon. The animals were released after the veterinary surgeon considered that they had fully recovered from the anesthesia. The fitted collars provided the identity of the wearer, its location, its elevation, and measurements related to the accuracy of the collar (e.g., degree of precision—see Coelho et al. 2007). Field tests of these collars have shown them to be much more accurate in the field than conventional very-high-frequency radiotracking collars (Coelho et al. 2007). We programmed these collars to record the aforementioned data on a weekly basis: Tuesdays at 0100, 0700, 1300, and 1900 h; Thursdays at 0300, 0900, 1500, and 2100 h; and Saturdays at 0500, 1100, 1700, and 2300 h. The collars were programmed to spend up to 240 s procuring satellites; if they did not manage to procure a sufficient number of satellites (minimum of 3) in this time period, no data were recorded. Data with large error values (> 15 m) were eliminated from the data set.

Because of a problem with the collar on the male, he was recaptured in December 2000 and a new collar was fitted; however, soon after this recapture the new collar developed the same problem, and consequently this limited our data collection on the male. In January 2001, the collar on the female also had a problem and so she was recaptured and a new collar fitted, which subsequently worked perfectly until the end of the study. Thus, we experienced a 50% failure rate in our GPS collars.

Geographic information system analyses.—The area of the reserve was converted into a rectangle measuring 20 × 18 km, centered on the geographic mean of all data points (fixes from the GPS collars), and satellite images of this area were acquired for analysis and interpretation. We decided to use an area larger than the home ranges for 2 reasons. First, our sampling frequency meant that the maned wolves could be using areas not recorded by the GPS collars; and 2nd, we needed to measure their potential as well as their actual habitat (Garshelis 2000). Our attempts to capture maned wolves indicated that the 2 individuals used in our study probably did not have neighboring pairs and their home range was therefore not socially limited by territorial behavior (Dietz 1984). The reserve is surround by high mountains (>2,200 m in elevation) and these seemed to be a spatially limiting factor, which resulted in us creating the aforementioned rectangle rather than a square.

The analysis and interpretation of satellite images from the years 2000 and 2002 was done using the software ArcGIS version 8.2 (Environmental Systems Research Institute 2002). this software permitted us to identify the following information for each fix (location used by maned wolves) determined by the GPS tracking collars: vegetation type or land use, elevation, distance to human constructions (i.e., buildings), distance to roads and trails, and distance to water sources (lakes, rivers, and streams). To confirm the accuracy of these data, a botanist was sent to verify these data at each of the locations used by the maned wolves (i.e., we ground-truthed 100% of our data); no discrepancy was detected.

Statistical analysis.—Home-range data were calculated using the minimum convex polygon (MCP) method in the software Biotas version 1.3 (de Melo et al. 2007) and core area was calculated using 50% ellipses (Cooper et al. 2006).

Following the recommendations of Garshelis (2000), data were analyzed by individual because of possible sex differences in habitat selection. Furthermore, given that maned wolves are largely nocturnal, we separated the data into day and night. Finally, the environment where we studied maned wolves is highly seasonal; therefore, data were further subdivided into wet and dry season. During the course of the study, the female gave birth to a pup; we therefore separated these data for analysis. In total, data were grouped into 10 types: female day dry season, female night dry season, female day wet season, female night wet season, female day reproduction, female night reproduction, male day dry season, male night dry season, male day wet season, and male night wet season. The factors we analyzed in relation to these groups of data were vegetation type and land use (10 categories), proximity to buildings (7 categories each of 250 m, starting at 0–250 m), proximity to roads or trails (7 categories each of 250 m, starting at 0–250 m), elevation (5 categories each of 250 m, starting at 750−1,000 m), and proximity to water (4 categories each of 250 m, starting at 0–250 m). These data all were analyzed by chi-square tests, with our expected data being calculated from interpretation of the satellite image (Table 1). When a category had <5 counts in a cell, it was eliminated from the analysis (such low counts can invalidate the chi-square test—Siegel and Castellan 1988), therefore our degrees of freedom are not always the number of categories minus 1. In the case of statistically significant results (P < 0.05), we conducted residual analyses to determine where significant differences were occurring (Siegel and Castellan 1988). Positive residuals with values >2 show that an individual was associated with the factor more than expected by chance, whereas negative residuals with values <—2 indicate that an animal was significantly avoiding the factor.

Table 1

Vegetation cover and land use in our study area in a cerrado-Atlantic forest transition zone in Brazil, and their use by the female and male maned wolves (Chrysocyon brachyurus).a

 Study area Female (%) Male (%) 
Vegetation cover Area (km2% area Day wet Day dry Night wet Night dry Reprod. day Reprod. night Day wet Day dry Night wet Night dry 
Water 0.37 0.10 0.00 0.00 0.00 0.00 0.00 0.34 0.00 1.89 0.00 0.00 
Rain forest 208.43 57.26 20.41 23.26 35.00 43.48 7.79 25.84 24.24 28.30 41.58 45.61 
Candeia forest 19.81 5.44 1.02 2.33 6.00 0.00 0.00 2.01 1.01 3.77 0.99 8.77 
Rocky fields 46.16 12.68 54.08 51.16 44.00 41.30 76.62 60.40 48.48 49.06 45.54 28.07 
Meadows 15.76 4.33 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Farmland 4.75 1.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.75 
Bare soil 0.78 0.21 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.77 0.00 0.00 
Eucalyptus 4.32 1.19 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Rocky areas 59.51 16.35 24.49 23.26 15.00 15.22 15.58 11.41 26.26 13.21 11.88 15.79 
Mining area 4.11 1.13 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Total (nNR NR 98 43 100 46 231 298 99 53 101 57 
 Study area Female (%) Male (%) 
Vegetation cover Area (km2% area Day wet Day dry Night wet Night dry Reprod. day Reprod. night Day wet Day dry Night wet Night dry 
Water 0.37 0.10 0.00 0.00 0.00 0.00 0.00 0.34 0.00 1.89 0.00 0.00 
Rain forest 208.43 57.26 20.41 23.26 35.00 43.48 7.79 25.84 24.24 28.30 41.58 45.61 
Candeia forest 19.81 5.44 1.02 2.33 6.00 0.00 0.00 2.01 1.01 3.77 0.99 8.77 
Rocky fields 46.16 12.68 54.08 51.16 44.00 41.30 76.62 60.40 48.48 49.06 45.54 28.07 
Meadows 15.76 4.33 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Farmland 4.75 1.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.75 
Bare soil 0.78 0.21 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.77 0.00 0.00 
Eucalyptus 4.32 1.19 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Rocky areas 59.51 16.35 24.49 23.26 15.00 15.22 15.58 11.41 26.26 13.21 11.88 15.79 
Mining area 4.11 1.13 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 
Total (nNR NR 98 43 100 46 231 298 99 53 101 57 
a

Wet = wet season; dry = dry season; Reprod. = reproduction; NR = not relevant.

Results

Home range.—The home ranges of the male and female maned wolves, using all the available data for each individual and calculated using the minimum convex polygon method (MCP100) were 64.66 km2 for the male and 72.28 km2 for the female (total number of fixes used is shown in Table 1). Using the MCP method and 95% of the data (MCP95), the home range of the male reduced to 33.54 km2 and that of the female reduced to 32.93 km2. Sizes of the home ranges reached asymptotes at 155 and 220 fixes, for the male and female, respectively (however, see de Melo et al. 2007). The degree of overlap of their home ranges was >90%. A frequency eclipse analysis of core areas, using only 50% of the data, showed that the core areas of both individuals were concentrated around the point of supplemental feeding; however, the core areas were still relatively large (10.68 km2 and 7.64 km2, for the male and female, respectively). During her reproductive phase, which occurred during the dry season, the female reduced her home range to 45.48 km2 (MCP100) or 15.38 km2 (MCP95), and this area showed a 100% overlap with her home range when nonreproductive. The maximum distance between 2 points used by 1 individual was 17.83 km.

Vegetation.—The chi-square analyses showed that maned wolves were not using their environment in proportion to the vegetation types available (Table 2). The residual analyses showed that maned wolves were actively avoiding rain forest and using rocky fields disproportionately (Fig. 1). This avoidance of rain forest and use of rocky fields was particularly pronounced in the female during her reproductive phase (Table 2; Fig. 1).

Table 2

Results of chi-square tests for use of different vegetation cover or land use by maned wolves (Chrysocyon brachyurus).

Individual Season Day/night χ2 d.f. P 
Female Wet Day 168.98 2 <0.001 
Female Wet Night 83.94 <0.001 
Female Dry Day 68.09 <0.001 
Female Dry Night 29.55 <0.001 
Female Reproductive Day 853.87 <0.001 
Female Reproductive Night 592.35 <0.001 
Male Wet Day 120.41 <0.001 
Male Wet Night 89.65 <0.001 
Male Dry Day 58.77 <0.001 
Male Dry Night 14.39 <0.01 
Individual Season Day/night χ2 d.f. P 
Female Wet Day 168.98 2 <0.001 
Female Wet Night 83.94 <0.001 
Female Dry Day 68.09 <0.001 
Female Dry Night 29.55 <0.001 
Female Reproductive Day 853.87 <0.001 
Female Reproductive Night 592.35 <0.001 
Male Wet Day 120.41 <0.001 
Male Wet Night 89.65 <0.001 
Male Dry Day 58.77 <0.001 
Male Dry Night 14.39 <0.01 
Fig. 1

Standardized residual analyses of different habitat types used by maned wolves.

Fig. 1

Standardized residual analyses of different habitat types used by maned wolves.

Buildings.—The chi-square tests showed that the maned wolves were associating significantly with buildings (Table 3). Maned wolves were located disproportionately close to buildings during the night (<250 m) and were indifferent to buildings during the day. These effects were most pronounced during the wet season for both male and female at night (Table 3) and for the female during reproduction (Table 3). Standardized residuals were only significant for the distance category of 0−250 m and were positive in all cases.

Table 3

Results of chi-square tests for proximity of maned wolves (Chrysocyon brachyurus) to buildings.

Individual Season Day/night χ2 d.f. P 
Female Wet Day 40.86 <0.001 
Female Wet Night 873.07 <0.001 
Female Dry Day 7.05 <0.05 
Female Dry Night 423.57 <0.001 
Female Reproductive Day 17.46 <0.01 
Female Reproductive Night 885.43 <0.001 
Male Wet Day 115.18 <0.001 
Male Wet Night 1682.75 <0.001 
Male Dry Day 19.85 <0.01 
Male Dry Night 640.55 <0.001 
Individual Season Day/night χ2 d.f. P 
Female Wet Day 40.86 <0.001 
Female Wet Night 873.07 <0.001 
Female Dry Day 7.05 <0.05 
Female Dry Night 423.57 <0.001 
Female Reproductive Day 17.46 <0.01 
Female Reproductive Night 885.43 <0.001 
Male Wet Day 115.18 <0.001 
Male Wet Night 1682.75 <0.001 
Male Dry Day 19.85 <0.01 
Male Dry Night 640.55 <0.001 

Roads and trails.—The chi-square tests showed that the maned wolves were associating significantly with roads and trails (Table 4). Standardized residual analyses showed that during the nights the wolves were located disproportionately close to trails (<250 m), and both during the day and night they avoided distances > 1,000 m from roads and trails. However, during the day maned wolves were indifferent about distances < 1,000 m from trails.

Table 4

Results of chi-square tests for proximity of maned wolevs (Chrysocyon brachyurus) to roads and trails

Individual Season Day/night χ2 df P 
Female Wet Day 45.57 <0.001 
Female Wet Night 124.24 <0.001 
Female Dry Day 14.90 <0.05 
Female Dry Night 66.25 <0.001 
Female Reproductive Day 178.42 <0.001 
Female Reproductive Night 346.00 <0.001 
Male Wet Day 71.75 <0.001 
Male Wet Night 22.69 <0.01 
Male Dry Day 205.79 <0.001 
Male Dry Night 84.74 <0.001 
Individual Season Day/night χ2 df P 
Female Wet Day 45.57 <0.001 
Female Wet Night 124.24 <0.001 
Female Dry Day 14.90 <0.05 
Female Dry Night 66.25 <0.001 
Female Reproductive Day 178.42 <0.001 
Female Reproductive Night 346.00 <0.001 
Male Wet Day 71.75 <0.001 
Male Wet Night 22.69 <0.01 
Male Dry Day 205.79 <0.001 
Male Dry Night 84.74 <0.001 

Elevation.—The maned wolves showed a nonrandom use of elevation (Table 5). Standardized residual analyses showed a strong preference for elevations of 1,251−1,500 m, avoidance of elevations < 1,250 m, and indifference to elevations >1,501 m.

Table 5

Results of chi-square tests for association of maned wolves (Chrysocyon brachyurus) with elevation.

Individual Season Day/night χ2 df P 
Female Wet Day 194.46 <0.001 
Female Wet Night 189.26 <0.001 
Female Dry Day 72.73 <0.001 
Female Dry Night 77.76 <0.001 
Female Reproductive Day 414.05 <0.001 
Female Reproductive Night 573.85 <0.001 
Male Wet Day 205.39 <0.001 
Male Wet Night 207.97 <0.001 
Male Dry Day 71.91 <0.001 
Male Dry Night 85.21 <0.001 
Individual Season Day/night χ2 df P 
Female Wet Day 194.46 <0.001 
Female Wet Night 189.26 <0.001 
Female Dry Day 72.73 <0.001 
Female Dry Night 77.76 <0.001 
Female Reproductive Day 414.05 <0.001 
Female Reproductive Night 573.85 <0.001 
Male Wet Day 205.39 <0.001 
Male Wet Night 207.97 <0.001 
Male Dry Day 71.91 <0.001 
Male Dry Night 85.21 <0.001 

Hydrology.—The only statistically significant relationship with hydrology was for habitat use by the female during her reproductive phase (day: χ2 = 9.58, d.f. = 1, P < 0.01; night: χ2 = 6.15, d.f. = 1, P < 0.05). Standardized residual analyses showed that she avoided distances >250 m from water; however, her den site was within 250 m of water. The maximum distance from water by a maned wolf in this study was 625 m, with a maximum potential distance of 950 m.

Discussion

Examination of our data shows that even in an environment with abundant rain forest and a much smaller area of rocky fields, wild maned wolves disproportionately used the cerradotype habitat and avoided the rain forest. Furthermore, maned wolves associated strongly with environmental features that can benefit them; for example, the strong association with buildings was due to supplemental feeding occurring there, whereas the strong association with roads and trails was likely due to the ease of locomotion provided by these features. The strong preference for elevation seemed to be related to the elevation at the source of supplemental feeding, and the majority of the cerrado habitat also was at this elevation. Because of the high density of water sources in the reserve, it is perhaps not surprising that maned wolves showed little association with their presence except for the female during her reproductive phase.

Examination of our data demonstrates that wild maned wolves have a preference for cerrado-type habitats. The majority of studies of wild maned wolves have been conducted in open habitats such as cerrado where maned wolves are suggested to be more common (Bueno et al. 2002; Dietz 1984; Motta-Junior and Martins 2002; Motta-Junior et al. 1996). However, surveys of maned wolves have not been undertaken in closed habitats such as forests. This preference for open habitats is probably due to maned wolves being highly adapted to hunting small prey in open environments (i.e., they possess disproportionately long legs, acute vision, and acute hearing). However, rain forests, as indicated by litterfall, are more productive habitats than the cerrado (Nardoto et al. 2006). Food availability in the cerrado is greater during the dry season (Alho et al. 1986) than in the wet season, when availability of fruit and insects is normally greater in forests such as the Atlantic forest (Develey and Peres 2000). However, maned wolves are not anatomically designed to exploit this increased food availability in the rain forest; for example, they are unable to climb trees.

Preservation of cerrado habitat in Brazil will probably best meet the requirements of maned wolves for survival. Although the study of Santos et al. (2003) showed that maned wolves are able to use cattle pastures in former Atlantic forest habitat, that study focused primarily on food habits of maned wolves. Thus, although the fragmentation of the Atlantic forest due to converting land to agricultural use (specifically for the production of cattle) may have allowed the maned wolf to expand its distribution into the Atlantic forest zone in Brazil, this habitat may be suboptimal for maned wolves. Our study showed that maned wolves avoided rain forest and never used farmland, supporting this hypothesis. However, to properly test this hypothesis, a study of reproductive success of maned wolves in the 2 habitat types is needed. Presently, both the cerrado and the Atlantic forest are biodiversity hotspots under threat from agricultural exploitation (Myers et al. 2000) and in the long-term this may have adverse consequences for populations of maned wolves.

It should be noted that although many of our telemetry fixes occurred close to the point of supplemental feeding, the wolves travelled extensively within the reserve, as indicated by their home-range sizes, which are within the range of those found in other studies (21.7-115 km2Carvalho and Vasconcellos 1995; de Melo et al. 2007; Dietz 1984). The core areas of the maned wolves in our study also were considerable (7.64–10.68 km2). Furthermore, the supplemental feeding was not sufficient to eliminate all hunting behavior; a concurrent study on the diet of these animals showed that in addition to the supplemental food, they had a diet comparable to other maned wolves (Silva and Talamoni 2004). This study estimated that 70% of their diet came from natural sources. Therefore, we can conclude that although the supplemental feeding did modify some aspects of the behavior of the maned wolves in our study, it did not overwhelmingly modify behavior or habitat use (Cooper et al. 2006).

The association of maned wolves with buildings was a consequence of supplemental feeding occurring at this location. There have been various reports in the Brazilian media of wild maned wolves approaching urban areas or farmhouses to receive supplemental feeding. This strong affinity for buildings was most pronounced at night when maned wolves naturally would be hunting in the wild. Coincidently, the supplemental food during our study was usually made available during the hours of darkness. Although the ultimate impact of supplemental feeding of these wild maned wolves can be debated, it obviously allows individuals to have easy access to a high-quality food source. During supplemental feeding, the maned wolves were wary of the presence of humans (C. M. Coelho, in litt.), however, the Brazilian media have shown films of maned wolves entering homes to be fed. The reduction of natural habitat combined with the growth of Brazilian cities means that such contacts between people and wild animals will increase.

The association of maned wolves with roads and trails can be explained as a means by which wild maned wolves can save energy and travel a long distances rapidly. For example, caribou (Rangifer tarandus) preferentially use roads as migration routes, despite the fact that this results in collisions with vehicles and increased wolf predation (Trombulak and Frissell 2000). In our study site, there have been no collisions reported with maned wolves on roads. On the public highway system in Brazil, however, maned wolves are victims of collisions (C. M. Coelho, in litt.), although this problem has not received proper quantification. Trail systems obviously do not pose risks for collisions with vehicles and, therefore, offer an efficient route for maned wolves to travel. As for their association with buildings, examination of our data shows a much more pronounced association with roads and trails at night, which coincides with natural hunting activity of wild maned wolves (de Melo et al. 2007; Sábato et al. 2006).

In our study, elevation, the site where supplemental feeding occurred, and vegetation type all were confounded. The preferred rocky fields and hotel both occurred at an elevation of 1,250–1,500 m; it is therefore not possible to determine whether maned wolves had a preferred elevation. However, our results did show that maned wolves avoided elevations lower than 1,250 m, but again this seems to be due to nonpreferred vegetation cover and land use at these lower elevations (e.g., bare soil, eucalyptus, and mining area). In other locations within Brazil, maned wolves can be found at lower elevations when their preferred vegetation types (i.e., cerrado) occur at these elevations. However, the rocky fields (campo rupestre) preferred in this study normally occur on the tops of high tablelands in the Brazilian cerrado biome at altitudes between 1,000 m and 1,800 m (Elten 1992). The indifference of maned wolves to elevations > 1,500 m is likely an artifact of low cell counts in the chi-square tests rather than real indifference. This was demonstrated by their avoidance of rocky areas, which occurred mainly at higher elevations.

The high density of water sources in our study area meant that it was difficult to determine the importance of this resource to maned wolves. For example, 84.17% of the reserve was within 250 m of a water source. The one interesting result was that when the female gave birth, she located her den <250 m from a water source. This perhaps reflected the need of a lactating female for water.

Time of day had a greater influence than did season on most of the associations we detected. It was not possible to separate the influence of reproduction from that of season. Given that maned wolves normally hunt at night (de Melo et al. 2007; Dietz 1984) and are unable to take advantage of the greater availability of fruit and insects in the Atlantic forest during the wet season (Develey and Peres 2000), it is not surprising that time of day was a more important factor than season on the variables that we measured. Outside of the reproductive season, the effect of sex on the variables measured seemed to be negligible. Unfortunately, because of equipment failure, we do not have data on the male during the reproductive phase, when it is most likely that differences in data from the male and female would be detected (de Melo et al. 2007). During her reproductive phase, the female increased use of rocky fields and showed greater association with buildings and roads and trails, which are all probably a result of her greater need for energy during lactation. The increase in use of rocky fields probably reflected increased hunting opportunities, the association with buildings reflected the availability of supplemental food, and the use of roads and trails provided a means to reduce costs of locomotion and reduce time away from the pup.

We have only informal observations on the behaviors performed in different habitat types. Maned wolves used the cerrado for hunting at night, and for reproduction; forest was used sometimes during the day for resting and at night the maned wolves were seen moving through the rain forest. Thus, although maned wolves can be found in a variety of environments in Brazil, examination of our data shows that their classic habitat, the cerrado, seems to be the most important habitat to emphasize for their conservation. However, in terms of survival and reproduction, this conclusion needs to be confirmed by an experiment designed to examine demographic responses to habitat (Garshelis 2000).

Acknowledgments

We thank Padres Célio and Sebastião and Consuelo from Caraça Private Natural Heritage Reserve for giving us permission to conduct our study in their reserve. The research in this paper was generously funded by Companhia Energética de Minas Gerais (energy company of Minas Gerais), and was conducted in compliance with all relevant state and federal laws in Brazil. We also wish to thank Instituto Brasileiro do Meio Ambiente e dos Recursos for the license to conduct this research. Finally, we are grateful to H. Buchanan-Smith, University of Stirling, for her comments on this manuscript.

Literature Clted

Alho
C. J. R.
Pereira
L. A.
Paula
A. C.
.
1986
.
Patterns of habitat utilization by small mammal populations in cerrado biome of central Brazil
.
Mammalia
 
50
:
447
460
.
Broomhall
L. S.
Mills
L. G. M.
du Toit
J. T.
.
2003
.
Home range and habitat use by cheetahs (Acinonyx jubatus) in the Kruger National Park
.
Journal of Zoology (London)
 
261
:
119
128
.
Bueno
A. A.
Belentani
S. C. S.
Motta-Junior
J. C.
.
2002
.
Feeding ecology of the maned wolf, Chrysocyon brachyurus (Illiger, 1815) (Mammalia: Canidae), in the ecological station of Itirapina, São Paulo State, Brazil
.
Biota Neotropica
 
2
:
1
9
.
Carvalho
C. T.
Vasconcellos
L. E. M.
.
1995
.
Disease, food and reproduction of the maned wolf Chrysocyon brachyurus (Illiger) (Carnivora, Canidae) in southeast Brazil
.
Revista Brasileira de Zoologia
 
12
:
627
640
.
Coelho
C. M.
de Melo
L. F. B.
Sábato
M. A. L.
Nogueira
D. R.
Young
R. J.
.
2007
.
A note on the use of GPS collars to monitor wild maned wolves Chrysocyon brachyurus (Illiger 1815) (Mammalia, Canidae)
.
Applied Animal Behaviour Science
 
105
:
259
264
.
Cooper
S. M.
Owens
M. K.
Cooper
R. M.
Ginnett
T. F.
.
2006
.
Effects of supplemental feeding on spatial distribution and browse utilization by white-tailed deer in semi-arid rangeland
.
Journal of Arid Environments
 
66
:
716
726
.
de Melo
L. F. B.
SÁbato
M. A. L.
Magni
E. M. V.
Young
R. J.
Coelho
C. M.
.
2007
.
The secret lives of maned wolves Chrysocyon brachyurus Illiger 1815: as revealed by GPS tracking collars
.
Journal of Zoology (London)
 
271
:
27
37
.
Develey
P. F.
Peres
C. A.
.
2000
.
Resource seasonality and the structure of mixed species bird flocks in a costal Atlantic forest of southeastern Brazil
.
Journal of Tropical Ecology
 
16
:
33
53
.
Dietz
J. M.
1984
.
Ecology and social organization of the maned wolf (Chrysocyon brachyurus)
.
Smithsonian Contributions to Zoology
 
392
:
1
51
.
Dietz
J. M.
1985
.
Chrysocyon brachyurus
.
Mammalian Species
 
234
:
1
4
.
Elten
G.
1992
.
Natural Brazilian vegetation types and their consequences
.
Anais da Academia Brasileira de Ciências
 
64
:
35
65
.
Environmental Systems Research Institute
.
2002
.
Arc View GIS v. 8.2
 . http://www.esri.com/data/index.html.
Accessed 20 February 2003
.
Garshelis
D. L.
2000
.
Delusions in habitat evaluation: measuring use, selection, and importance
. Pp.
111
164
in
Research techniques in animal ecology: controversies and consequences
  (
Boitani
L.
Fuller
T. K.
, eds.).
Columbia University Press
,
New York
.
Giulietti
A. M.
Pirani
J. R.
.
1988
.
Patterns of geographic distribution of some plant species from the Espinhaço Range, Minas Gerais and Bahia, Brazil
. Pp.
39
69
in
Proceedings of a workshop on neotropical distribution patterns
  (
Vanzolini
P. E.
Heyer
W. R.
, eds.).
Academia Brasileira de Ciências
,
Rio de Janeiro, Brazil
.
Giulietti
A. M.
Pirani
J. R.
Harley
R. M.
.
1997
.
Espinhaço Range region, eastern Brazil
. Pp.
397
404
in
Centres of plant diversity, a guide and strategy for their conservation
  (
Davis
S. D.
Heywood
V. H.
Herrera-MacBryde
O.
Villa-Lobos
J.
Hamilton
A. C.
, eds.).
Information Press
,
Oxford, United Kingdom
.
Hulme
P. E.
2006
.
Beyond control: wider implications for the management of biological invasions
.
Journal of Applied Ecology
 
43
:
835
847
.
Motta-Junior
J. C.
Martins
K.
.
2002
.
The frugivorous diet of the maned wolf, Chrysocyon brachyurus, in Brazil: ecology and conservation
. Pp.
291
303
in
Seed dispersal and frugivory: ecology, evolution and conservation
  (
Levey
D. J.
Silva
W. R.
Galetti
M.
, eds.).
CABI Publishing
,
Wallingford, Oxfordshire, United Kingdom
.
Motta-Junior
J. C.
Talamoni
S. A.
Lombardi
J. A.
Simolomaki
K.
.
1996
.
Diet of the maned wolf, Chrysocyon brachyurus, in central Brazil
.
Journal of Zoology (London)
 
240
:
277
284
.
Myers
N.
Mittermeier
R. A.
Mittermeier
C. G.
da Fonseca
G. A. B.
Kent
J.
.
2000
.
Biodiversity hotspots for conservation priorities
.
Nature
 
403
:
853
858
.
Nardoto
G. B.
Bustamante
M. M. C.
Pinto
A. S.
Klink
C. A.
.
2006
.
Nutrient use efficiency at ecosystem and species level in savanna areas of central Brazil and impacts of fire
.
Journal of Tropical Ecology
 
22
:
191
201
.
Sábato
M. A. L.
de Melo
L. F. B.
Magni
E. M. V.
Young
R. J.
Coelho
C. M.
.
2006
.
A note on the effect of the full moon on the activity of wild maned wolves, Chrysocyon brachyurus
.
Behavioural Processes
 
73
:
228
230
.
Santos
E. F.
Setz
E. Z. F.
Gobbi
N.
.
2003
.
Diet of the maned wolf (Chrysocyon brachyurus) and its role in seed dispersal on a cattle ranch in Brazil
.
Journal of Zoology (London)
 
260
:
203
208
.
Siegel
S.
Castellan
N. J.
Jr.
1988
.
Nonparameteric statistics for the behavioral sciences
 .
2
nd ed.
McGraw Hill
,
Boston, Massachusetts
.
Silva
J. A.
Talamoni
S. A.
.
2004
.
Core area and centre of activity of maned wolves, Chrysocyon brachyurus (Illiger) (Mammalia, Canidae), submitted to supplemental feeding
.
Revista Brasileira de Zoologia
 
21
:
391
395
.
Tabarelli
M.
Pinto
L. P.
Silva
J. M. C.
Hirota
M.
Bede
L.
.
2005
.
Challenges and opportunities for biodiversity conservation in the Brazilian Atlantic forest
.
Conservation Biology
 
19
:
695
700
.
Teixeira
C. P.
Hirsch
A.
Perini
H.
Young
R. J.
.
2006
.
Marsupials from space: fluctuating asymmetry, geographical information systems and animal conservation
.
Proceedings of the Royal Society of London, B. Biological Sciences
 
273
:
1007
1012
.
Theuerkauf
J.
Jedrzejewski
W.
.
2002
.
Accuracy of radiotelemetry to estimate wolf activity and locations
.
Journal of Wildlife Management
 
66
:
859
864
.
Trombulak
S. C.
Frissell
C. A.
.
2000
.
Review of ecological effects of roads on terrestrial and aquatic communities
.
Conservation Biology
 
14
:
18
30
.
Williams
N. S. G.
McDonnell
M. J.
Phelan
G. K.
Keim
L. D.
van der Ree
R.
.
2006
.
Range expansion due to urbanization: increased food sources attract grey-head flying-foxes (Pteropus poliocephalus) to Melbourne
.
Austral Ecology
 
31
:
190
198
.
Wing
M. G.
Eklund
A.
Kellogg
L. D.
.
2005
.
Consumer-grade global positioning system (GPS) accuracy and reliability
.
Journal of Forestry
 
103
:
169
173
.

Author notes

Associate Editor was Rodrigo A. Medellín.